Whamcloud - gitweb
Chris Horn [Thu, 12 Aug 2021 16:26:07 +0000 (11:26 -0500)]
LU-14939 lnet: Allow specifying a source NID for lnetctl ping
Add a new --source option for lnetctl ping command. This allows the
user to specify a local NI from which to send the ping. This also
ensures that the specified destination NID is also used. Otherwise,
pings to multi-rail peers may end up going to a different peer NI
based on the multi-rail selection algorithm. The ability to specify
a source NI, and thus fix the destination NI, is a great help in
troubleshooting communication issues between multi-rail peers.
Add test to exercise lnetctl ping --source option.
HPE-bug-id: LUS-10296
Test-Parameters: trivial testlist=sanity-lnet
Signed-off-by: Chris Horn <chris.horn@hpe.com>
Change-Id: I454217b30a92414de537880f076a11a693b1f0b3
Reviewed-on: https://review.whamcloud.com/44727
Reviewed-by: Serguei Smirnov <ssmirnov@whamcloud.com>
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Andriy Skulysh <andriy.skulysh@hpe.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Sergey Cheremencev [Thu, 10 Jun 2021 10:08:28 +0000 (13:08 +0300)]
LU-14867 doc: a few words about an asterisk in lfs quota
Clarify the difference between an asterisk printed per OST
in verbouse lfs quota output and an asterisk near the whole
filesystem usage.
Change-Id: I778fe1f7b1f6f8d55c311d81bd2b311d82463390
Test-Parameters: trivial
Signed-off-by: Sergey Cheremencev <sergey.cheremencev@hpe.com>
Reviewed-on: https://review.whamcloud.com/44357
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-by: Petros Koutoupis <petros.koutoupis@hpe.com>
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Shaun Tancheff [Sat, 9 Oct 2021 04:33:23 +0000 (11:33 +0700)]
LU-14437 gnilnd: use ktime_get_seconds() to get time
Use ktime_get_seconds() to directly get the time inatead of
getting a timespec and converting it.
Fixes:
4b0e495e3c ("LU-14080 gnilnd: updates for SUSE 15 SP2")
Test-Parameters: trivial
Signed-off-by: Shaun Tancheff <shaun.tancheff@hpe.com>
Change-Id: I256855ceb9e038a9960fa76fe6e3bfe63fb16580
Reviewed-on: https://review.whamcloud.com/41679
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-by: James Simmons <jsimmons@infradead.org>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Artem Blagodarenko [Sat, 5 Dec 2020 01:40:52 +0000 (20:40 -0500)]
LU-14188 osd: remove not used iam_container lock
There is an rw_semaphore
struct iam_container {
...
/*
* read-write lock protecting index consistency.
*/
struct rw_semaphore ic_sem; <<<<<<
struct dynlock ic_tree_lock;
/* Protect ic_idle_bh */
struct mutex ic_idle_mutex;
...
};
There is initialization
2 234 lustre/osd-ldiskfs/osd_iam.c <<iam_container_init>>
init_rwsem(&c->ic_sem);
There are wrappers
3 622 lustre/osd-ldiskfs/osd_iam.c <<iam_container_write_lock>>
down_write(&ic->ic_sem);
4 627 lustre/osd-ldiskfs/osd_iam.c <<iam_container_write_unlock>>
up_write(&ic->ic_sem);
5 632 lustre/osd-ldiskfs/osd_iam.c <<iam_container_read_lock>>
down_read(&ic->ic_sem);
6 637 lustre/osd-ldiskfs/osd_iam.c <<iam_container_read_unlock>>
up_read(&ic->ic_sem);
But this wrappers are not used. And, based on the git history, never been used.
Let's delete this useless code.
Change-Id: Ied1122f034e53fee08888e1091f700bda4507f00
Signed-off-by: Artem Blagodarenko <artem.blagodarenko@hpe.com>
HPE-bug-id: LUS-9545
Reviewed-on: https://review.whamcloud.com/40890
Reviewed-by: Neil Brown <neilb@suse.de>
Reviewed-by: Alex Zhuravlev <bzzz@whamcloud.com>
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Mr NeilBrown [Tue, 9 Nov 2021 15:55:33 +0000 (10:55 -0500)]
LU-14073 ldiskfs: update ldiskfs patches for Linux 5.9
Minor conflict changes only.
Test-Parameters: trivial
Signed-off-by: Mr NeilBrown <neilb@suse.de>
Change-Id: If0b58fc278f6721e44e26c6052509c31068f8e78
Reviewed-on: https://review.whamcloud.com/40397
Reviewed-by: James Simmons <jsimmons@infradead.org>
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Alex Zhuravlev <bzzz@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
James Simmons [Mon, 1 Nov 2021 18:31:24 +0000 (14:31 -0400)]
LU-9325 obdclass: make niduuid for lustre_stop_mgc() static
The process to create a proper string for niduuid can be made
simpler and avoid a memory allocation.
Change-Id: I52cb01117e41cbcf2756477e91934a42d31fd157
Signed-off-by: James Simmons <jsimmons@infradead.org>
Reviewed-on: https://review.whamcloud.com/33617
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Arshad Hussain <arshad.hussain@aeoncomputing.com>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-by: Alex Zhuravlev <bzzz@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Mr. NeilBrown [Tue, 16 Nov 2021 16:11:22 +0000 (11:11 -0500)]
LU-12511 ldlm: free resource when ldlm_lock_create() fails.
ldlm_lock_create() gets a resource, but don't put it on
all failure paths. It should.
Change-Id: Ib49bcafdeac834c412adad9db135034d1ea06a04
Signed-off-by: Mr. NeilBrown <neilb@suse.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Reviewed-on: https://review.whamcloud.com/45585
Tested-by: jenkins <devops@whamcloud.com>
Reviewed-by: Olaf Faaland-LLNL <faaland1@llnl.gov>
Reviewed-by: Andriy Skulysh <andriy.skulysh@hpe.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Mr. NeilBrown [Mon, 15 Nov 2021 19:16:54 +0000 (14:16 -0500)]
LU-12511 llite: fix misuse of current->parent.
current->parent is used by ptrace to redirect some signal delivery
to the ptracer. It should only be used by 'ptrace' or 'signal' code.
All other users should use current->real_parent, which is the real
parent.
Change-Id: I3212fd9f3db9935b8144dba8af82eb2f387a5c45
Signed-off-by: Mr. NeilBrown <neilb@suse.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Reviewed-on: https://review.whamcloud.com/45580
Tested-by: jenkins <devops@whamcloud.com>
Reviewed-by: Patrick Farrell <pfarrell@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Li Dongyang [Fri, 12 Nov 2021 12:24:05 +0000 (23:24 +1100)]
LU-15083 ldiskfs: disable xattr credits check
Add ext4-xattr-disable-credits-check.patch to the
ldiskfs series from kernel 4.15+ except the
rhel8 ones. They already have this fix.
Change-Id: I1f9818c394377cdba6b11d2b24c26d2354f41ac0
Test-Parameters: trivial
Signed-off-by: Li Dongyang <dongyangli@ddn.com>
Reviewed-on: https://review.whamcloud.com/45546
Tested-by: jenkins <devops@whamcloud.com>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Tested-by: Shuichi Ihara <sihara@ddn.com>
Tested-by: Oleg Drokin <green@whamcloud.com>
Reviewed-by: James Simmons <jsimmons@infradead.org>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Lei Feng [Thu, 11 Nov 2021 06:06:39 +0000 (14:06 +0800)]
LU-9162 lod: option to set max stripe count per filesystem
Add an option to set max default stripe count when the stripe count
is set to -1.
Change-Id: I02634a02a6f6579750fe964662b7e644af1689d6
Signed-off-by: Lei Feng <flei@whamcloud.com>
Test-Parameters: trivial
Reviewed-on: https://review.whamcloud.com/45532
Tested-by: jenkins <devops@whamcloud.com>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Li Xi <lixi@ddn.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Bobi Jam [Tue, 2 Nov 2021 07:27:59 +0000 (15:27 +0800)]
LU-14781 osp: osp_object_free access NULL pointer
If a osp_object is created by multiple threads at the same time,
lu_object_find_at() could allocate an osp_object without object
initialization, before hash inserting of the object, it find another
object has been created and inserted by another thread, it will free
the unintialized osp_object, and osp_object_free() will access
an uninitialized list_head (opo_xattr_list).
This patch initialize osp_object::opo_xattr_list in its allocation
function.
Call trace:
lu_object_free.isra.30+0xf2/0x170 [obdclass]
lu_object_find_at+0x496/0x930 [obdclass]
lod_initialize_objects+0x3e4/0xba0 [lod]
lod_parse_striping+0x693/0xc20 [lod]
lod_striping_load+0x2b2/0x660 [lod]
lod_declare_destroy+0x12b/0x600 [lod]
mdd_declare_finish_unlink+0x91/0x210 [mdd]
mdd_unlink+0x48f/0xab0 [mdd]
mdt_reint_unlink+0xc32/0x1550 [mdt]
mdt_reint_rec+0x83/0x210 [mdt]
mdt_reint_internal+0x6e1/0xb00 [mdt]
mdt_reint+0x67/0x140 [mdt]
tgt_request_handle+0xaee/0x15f0 [ptlrpc]
ptlrpc_server_handle_request+0x24b/0xab0 [ptlrpc]
ptlrpc_main+0xb34/0x1470 [ptlrpc]
kthread+0xd1/0xe0
Signed-off-by: Bobi Jam <bobijam@whamcloud.com>
Change-Id: Ib86aca5b41e94a1758f177655ea3a0f680335e0f
Reviewed-on: https://review.whamcloud.com/45442
Tested-by: jenkins <devops@whamcloud.com>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Alex Zhuravlev <bzzz@whamcloud.com>
Alexander Zarochentsev [Tue, 31 Aug 2021 05:36:30 +0000 (08:36 +0300)]
LU-15114 osp: changes queuing throttle
Prevent queue of sync changes from growing too much
by adding resends when queue size reaches
some (tunable) limit.
HPE-bug-id: LUS-10345
Signed-off-by: Alexander Zarochentsev <alexander.zarochentsev@hpe.com>
Change-Id: I5efabb91d3700c58d9451f81c5fed9a22ae404fb
Reviewed-on: https://review.whamcloud.com/45265
Reviewed-by: Alexander Boyko <alexander.boyko@hpe.com>
Reviewed-by: Andrew Perepechko <andrew.perepechko@hpe.com>
Reviewed-by: Lai Siyao <lai.siyao@whamcloud.com>
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Andrew Perepechko [Fri, 8 Oct 2021 10:05:50 +0000 (17:05 +0700)]
LU-14491 ldiskfs: do not corrupt journal with bh change rh7.6
Currently, ldiskfs_xattr_delete_inode() zeroes xattr inode
references in cached buffers that haven't been prepared by
get_write_access().
When using journal checksums, it is possible that these buffers
are modified after the checksum is calculated but before the
buffer has been written to journal. Journal replay will fail
with a journal checksum error message if this transaction needs
to be replayed.
This is a port of:
Lustre-commit:
d563f5140ffa183d0854cf7cd493ad884e314e3d
Lustre-change: https://review.whamcloud.com/41896
Test-Parameters: trivial
HPE-bug-id: LUS-9483
Signed-off-by: Shaun Tancheff <shaun.tancheff@hpe.com>
Change-Id: I5ce1016737a16ddf74811c43ae74296d4f3e03b0
Reviewed-on: https://review.whamcloud.com/45164
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Andrew Perepechko <andrew.perepechko@hpe.com>
Reviewed-by: Alexander Zarochentsev <alexander.zarochentsev@hpe.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Chris Horn [Tue, 5 Oct 2021 19:11:29 +0000 (14:11 -0500)]
LU-15068 ptlrpc: Do not unlink difficult reply until sent
If a difficult reply is queued in LNet, or the PUT for it is
otherwise delayed, then it is possible for the commit callback
to unlink the reply MD which will abort the send. This results in
client hitting "slow reply" timeout for the associated RPC and
an unnecessary reconnect (and possibly resend).
This patch replaces the rs_on_net flag with rs_sent and rs_unlinked.
These flags indicate whether the send event for the reply MD has
been generated, and whether the MD has been unlinked, respectively.
If rs_sent is set, but rs_unlinked has not been set, then ptlrpc_hr
is free to unlink the reply MD as a result of the commit callback.
The reply-ack will simply be dropped by the server.
If ptlrpc_hr is processing the reply because of commit callback, and
rs_sent has not been set, then ptlrpc_hr will not unlink the reply
MD. This means that the reply_out_callback must also be modified to
check for this case when the send event occurs. Otherwise, if the ACK
never arrives from the client, then the MD would never be unlinked.
Thus when the send event occurs, and rs_handled is set, the
reply_out_callback will schedule the reply for handling by ptlrpc_hr.
HPE-bug-id: LUS-10505
Signed-off-by: Chris Horn <chris.horn@hpe.com>
Change-Id: Ib8f4853c7ab35d72624fce7ee3fba9e59a746e1f
Reviewed-on: https://review.whamcloud.com/45138
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Andriy Skulysh <andriy.skulysh@hpe.com>
Reviewed-by: Alexey Lyashkov <alexey.lyashkov@hpe.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Sergey Cheremencev [Wed, 15 Sep 2021 15:05:45 +0000 (18:05 +0300)]
LU-15065 quota: fix BIO write performance drop
Before the patch qti_lqes_qunit_min used int to store qunit
value, while lqe_qunit type is _u64. lqe_qunit > 2G caused
an overflow in a local integer argument. For example, when
block hard limit was set to 500TB(i.e. lqe_qunit was about
64TB in a system with 2 OSTs), qti_lqes_qunit_min returned
0 instead of 64TB in a qmt_lvbo_fill. Thus new qunit was not
set on OSTs(qsd_set_qunit wasn't called). Without the qunit,
OST began to send release request after each acquire. For
example, to write 10MB at the OST were sent 2 acquire and
2 release reuests(as qunit was not set on OST). With the
fix, i.e. in a normal case, OST needs just one acquire
request. The issue caused performance drop in a bufferred
write up to 15%-20% if compare with a baseline without PQ
patches.
Note, the issue exists only when a hard limit is set to some
high value(>100GB). The exact hard limit value depends on OSTs
number in a system and on amount of used space, but let's think
that issue doesn't exist on a clean system with 2 OSTs and hard
block limit 100G(this case was checked).
Remove qmt_pool_hash - it is not used anywhere since
"LU-11023 quota: remove quota pool ID".
HPE-bug-id: LUS-10250
Change-Id: I2c4ce38f5b9395ed1f4868d4c8efc00751116b15
Signed-off-by: Sergey Cheremencev <sergey.cheremencev@hpe.com>
Reviewed-on: https://review.whamcloud.com/45133
Reviewed-by: Petros Koutoupis <petros.koutoupis@hpe.com>
Reviewed-by: Alexander Boyko <alexander.boyko@hpe.com>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Sergey Cheremencev [Fri, 29 Oct 2021 19:00:46 +0000 (22:00 +0300)]
LU-15179 tests: add trap cleanup_quota_test
Add stack_trap cleanup_quota_test to the tests that
use setup_quota_test. If a test fails without calling
cleanup_quota_test, it may cause later tests to fail
due to used space > 0.
Remove ${tdir}_dom, if exists, in cleanup_quota_test.
sanity-quota_75 doesn't remove test_dom directory.
Test-Parameters: trivial testlist=sanity-quota
Fixes:
a4fbe734("LU-14739 quota: nodemap squashed root cannot bypass quota")
Change-Id: Ife4fd499b427bee79f74a5e172d233fe6a83e240
Signed-off-by: Sergey Cheremencev <sergey.cheremencev@hpe.com>
Reviewed-on: https://review.whamcloud.com/45418
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-by: Alexander Zarochentsev <alexander.zarochentsev@hpe.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Sergey Cheremencev [Fri, 18 Jun 2021 16:08:16 +0000 (19:08 +0300)]
LU-15025 quota: stale edquot after clearing limits
When hard and soft limit set to 0, lqe enforced flag is also
set to false. As qmt_adjust_qunit does not handle not enforced
lqes, edquot set to the pool continues to be true and a user
gets -EDQUOT even if all pool limits are cleared. This was ok
for global pool lqe as since it turned off, zero limits are
sent to OSTs causing OSTs to release all granted space and
avoid EDQUOT. Fix this for PQ - set edquot and qunit to zero,
since appropriate lqe becomes "not enforced".
HPE-bug-id: LUS-10146
Change-Id: I1e08929bae7e1b37b1e8cbbc44859a786b5fb090
Reviewed-on: https://es-gerrit.dev.cray.com/158915/
Reviewed-by: Alexander Boyko <alexander.boyko@hpe.com>
Tested-by: Elena Gryaznova <c17455@cray.com>
Tested-by: Jenkins Build User <nssreleng@cray.com>
Signed-off-by: Sergey Cheremencev <sergey.cheremencev@hpe.com>
Reviewed-on: https://review.whamcloud.com/45000
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Shaun Tancheff <shaun.tancheff@hpe.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Sergey Cheremencev [Tue, 12 Oct 2021 15:21:49 +0000 (18:21 +0300)]
LU-15191 quota: set correct revoke_time
When we do qmt_adjust_qunit, there are several lqes
and lqe_revoke_time is set for some of them, it means
appropriate OSTs have been already notified with the
least qunit and there is no chance to free more space.
If a qunit of the current lqe becomes equal to the least
qunit, find an lqe with the minimum(earliest) revoke_time
and set this revoke_time to the current one.
This patch fixes the following case. For example, we have
8 OSTs and 4 MDTs(i.e. 12 slaves) and a pool with just one
OST. Global hard block limit for the user is 50M, and 10M
for this user in a pool. User's usage is 0. As global pool
has 12 slaves it's initial qunit value is 1M, i.e. equal to
the least qunit. At the same time initial qunit value for the
pool with one OST is 4M. When user begins to write, pool's
qunit is decreased to 1M, but lqe_revoke is not set - it
should be set only after sending new qunit to OSTs in
qmt_lvbo_update. However, it won't be send because appropriate
lge_qunit in lqe global array already has the same value.
This problem caused sanity-quota_72 to hang instead of fail
with EDQUOT in test_1_check_write.
HPE-bug-id: LUS-10516
Change-Id: I5878c1e719ae83a69ad5dbc3653717bb1b4de632
Reviewed-by: Andriy Skulysh <c17819@cray.com>
Reviewed-by: Alexander Boyko <alexander.boyko@hpe.com>
Tested-by: Elena Gryaznova <c17455@cray.com>
Signed-off-by: Sergey Cheremencev <sergey.cheremencev@hpe.com>
Reviewed-on: https://review.whamcloud.com/45447
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Andriy Skulysh <andriy.skulysh@hpe.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
James Simmons [Fri, 3 Sep 2021 23:22:18 +0000 (19:22 -0400)]
LU-10391 uapi: move out kernel only code.
Userland doesn't use the new IPv6 function in the UAPI nidstr.h.
So move this to the kernel header lib-types.h. Normally this
wouldn't matter but the python wrappers break with the LUTF
project. With the move to Netlink in the future for the user land
API we shouldn't need these functions anyways.
Test-parameters: trivial
Change-Id: I2ac6d102d4575f639573253c21723d62ca08abc2
Signed-off-by: James Simmons <jsimmons@infradead.org>
Reviewed-on: https://review.whamcloud.com/44844
Reviewed-by: Neil Brown <neilb@suse.de>
Reviewed-by: Chris Horn <chris.horn@hpe.com>
Reviewed-by: Amir Shehata <ashehata@whamcloud.com>
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Vladimir Saveliev [Thu, 9 Sep 2021 12:05:24 +0000 (15:05 +0300)]
LU-12347 llite: do not take mod rpc slot for getxattr
The following scenario may lead to client eviction:
clientA clientB MDS
threadA1: write to file F1, get
and hold DoM MDC LDLM lock L1:
->cl_io_loop()
->cl_io_lock()
:
->mdc_lock_granted()
->lock->l_writers++
[hold ref until write done]
threadA2-A8: create files F2-F8:
->ll_file_open()
->mdc_enqueue_base()
->ldlm_cli_enqueue()
->ptlrpc_get_mod_rpc_slot()
->ptlrpc_queue_wait()
[hold RPC slot until create done]
OST(s) in recovery.
MDS waiting on OST(s) to
precreate new objects.
threadA1:
-> cl_io_start()
-> __generic_file_aio_write()
-> file_remove_suid()
-> ll_xattr_cache_refill()
-> mdc_xattr_common()
-> ptlrpc_get_mod_rpc_slot()
[blocked waiting for RPC slot]
threadB1: write file F1,
enqueue DoM MDC lock L1
MDS sends blocking AST
to clientA for lock L1
ldlm_threadA3: cannot cancel busy lock L1:
-> ldlm_handle_bl_callback()
["Lock L1 referenced, will be cancelled later"]
MDS evicts clientA for
not cancelling lock L1
threadA1: never completes write:
->cl_io_end()
->cl_io_unlock()
->osc_lock_cancel()
->lock->l_writers--;
The fix is to add IT_GETXATTR to list of operations which do not
need mod rpc slot.
Tests to illustrate the issue is added.
wait_for_function(): total sleep time (wait) is to be equal to max
when 1 is returned.
Signed-off-by: Vladimir Saveliev <vlaidimir.saveliev@hpe.com>
HPE-bug-id: LUS-7271
Change-Id: I1b80677df084bda141b9ac58a78b765bd0b14a41
Reviewed-on: https://review.whamcloud.com/44151
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-by: Andrew Perepechko <andrew.perepechko@hpe.com>
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Mr NeilBrown [Wed, 10 Feb 2021 22:49:54 +0000 (09:49 +1100)]
LU-14428 libcfs: separate daemon_list from cfs_trace_data
cfs_trace_data provides a fifo for trace messages. To minimize
locking, there is a separate fifo for each CPU, and even for different
interrupt levels per-cpu.
When a page is remove from the fifo to br written to a file, the page
is added to a "daemon_list". Trace message on the daemon_list have
already been logged to a file, but can be easily dumped to the console
when a bug occurs.
The daemon_list is always accessed from a single thread at a time, so
the per-CPU facilities for cfs_trace_data are not needed. However
daemon_list is currently managed per-cpu as part of cfs_trace_data.
This patch moves the daemon_list of pages out to a separate structure
- a simple linked list, protected by cfs_tracefile_sem.
Rather than using a 'cfs_trace_page' to hold linkage information and
content size, we use page->lru for linkage and page->private for
the size of the content in each page.
This is a step towards replacing cfs_trace_data with the Linux
ring_buffer which provides similar functionality with even less
locking.
In the current code, if the daemon which writes trace data to a file
cannot keep up with load, excess pages are moved to the daemon_list
temporarily before being discarded. With the patch, these page are
simply discarded immediately.
If the daemon thread cannot keep up, that is a configuration problem
and temporarily preserving a few pages is unlikely to really help.
Signed-off-by: Mr NeilBrown <neilb@suse.de>
Change-Id: Ie894f7751cadacb515215f18182163ea5d26e969
Reviewed-on: https://review.whamcloud.com/41493
Reviewed-by: James Simmons <jsimmons@infradead.org>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Andriy Skulysh [Wed, 16 Sep 2020 11:38:41 +0000 (14:38 +0300)]
LU-15107 mdt: Exclusive create isn't replayed
mdt_finish_open() fails on exclusive create check as
there isn't an infrormation that the file was created.
Set DISP_OPEN_CREATE for exclusive open replay,
as we know that the original request has succeeded.
Change-Id: Idc71db76b757670b91b3bb1718870a018a805ed2
HPE-bug-id: LUS-9358
Signed-off-by: Andriy Skulysh <c17819@cray.com>
Reviewed-by: Vitaly Fertman <c17818@cray.com>
Reviewed-by: Alexander Zarochentsev <c17826@cray.com>
Reviewed-on: https://review.whamcloud.com/45254
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-by: Alexander Zarochentsev <alexander.zarochentsev@hpe.com>
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
James Simmons [Wed, 27 Oct 2021 16:03:10 +0000 (12:03 -0400)]
LU-9680 net: Netlink improvements
With the expansion of the use of Netlink several issues have been
encountered. This patch fixes many of the issues. The issues are:
1) Fix idx handling in lnet_genl_parse_list() function. It needs
to always been incremented. Some renaming suggestion for
enum lnet_nl_scalar_attrs from Neil. New LN_SCALAR_ATTR_INT_VALUE
to allow pushing integers as well as strings from userspace.
2) Create struct genl_filter_list which will be used to create
a list of items to pass back to userland. This will be a common
setup.
3) A normal user can't read /sys/debug/kernel/lustre which breaks
lctl ***_params XXX since the first function called is
llapi_param_get_paths(). Without the ability to read the
debugfs tree glob() will fail. The solution is to use the
kernel's glob function and just pass the requested string to
the kernel.
4) For the external coordinator work you create a YAML parser
that listens for kernel generated Netlink packets. This is
a continuous stream vs an one time reply which we don't
handle correctly. We move the handling of the completion
of a Netlink packet series icompletely into the function
yaml_netlink_msg_complete. In yaml_netlink_msg_parse() for
the async case add "---" and in yaml_netlink_msg_complete()
add "..." to define the beginning and end of a YAML document.
5) We have 3 types of setups. For kernel generated events it is
possible to use just a YAML parser to listen for events. For
the normal request -> reply setup we need both a YAML emitter
and YAML parser. The last case is just sending commands to
the kernel which only needs a YAML emitter. It is possible for
that action to fail so we need to add handling for errors to
the YAML emitter. We keep error handling for the YAML parser
as well to handle the case of a stand along YAML parser listener.
6) Reworked the code that translates YAML to Netlink packets to
send to the kernel so the both key and value pairs are sent
seperately for the mapping case. This avoids dealing with
complex string parsing in the kernel.
7) Error message handling was incorrect. struct nlmsgerr msg field
is also the start of the nlattrs for the ext ack handling.
Test-Parameters: trivial
Change-Id: Ic8eee8fd0020b7a63565de6ef69f2c74bf4bdcd8
Signed-off-by: James Simmons <jsimmons@infradead.org>
Reviewed-on: https://review.whamcloud.com/44358
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Ben Evans <beevans@whamcloud.com>
Reviewed-by: Petros Koutoupis <petros.koutoupis@hpe.com>
Tested-by: jenkins <devops@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Bobi Jam [Thu, 2 Sep 2021 14:24:50 +0000 (22:24 +0800)]
LU-14713 llite: mend the trunc_sem_up_write()
The original lli_trunc_sem replace change (commit
e5914a61ac) fixed a
lock scenario:
t1 (page fault) t2 (dio read) t3 (truncate)
|- vm_mmap_pgoff() |- vvp_io_read_start() |- vvp_io_setattr
|- down_write(mmap_sem) |- down_read(trunc_sem) _start()
|- do_map() |- ll_direct_IO_impl()
|- vvp_io_fault_start |- ll_get_user_pages()
|- down_write(
|- down_read(mmap_sem) trunc_sem)
|- down_read(trunc_sem)
t1 waits for read semaphore of trunc_sem which is hindered by t3,
since t3 is waiting for the write semaphore while t2 take its read
semaphore,and t2 is waiting for mmap_sem which has been taken by t1,
and a deadlock ensues.
commit
e5914a61ac changes the down_read(trunc_sem) to
trunc_sem_down_read_nowait() in page fault path, to make it ignore
that there is a down_write(trunc_sem) waiting, just takes the read
semaphore if no writer has taken the semaphore, and breaks the
deadlock.
But there is a delicacy in using wake_up_var(), wake_up_var()->
__wake_up_bit()->waitqueue_active() locklessly test for waiters on the
queue, and if it's called without explicit smp_mb() it's possible for
the waitqueue_active() to ge hoisted before the condition store such
that we'll observe an empty wait list and the waiter might not
observe the condition, and the waiter won't get woke up whereafter.
Fixes:
e5914a61ac ("LU-12460 llite: replace lli_trunc_sem")
Change-Id: Ifdda2c1c8a4171466be1723923c136e84de8ce0e
Signed-off-by: Bobi Jam <bobijam@whamcloud.com>
Reviewed-on: https://review.whamcloud.com/43844
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-by: Neil Brown <neilb@suse.de>
Reviewed-by: Patrick Farrell <pfarrell@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Tested-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Bobi Jam [Tue, 12 Oct 2021 12:02:01 +0000 (20:02 +0800)]
LU-14971 test: align mirror_io resync implementation
Align the mirror_io resync implementation with
llapi_mirror_resync_many().
Signed-off-by: Bobi Jam <bobijam@whamcloud.com>
Change-Id: Icf11c4c2302f36fc0f9682e0a310058081e1214f
Reviewed-on: https://review.whamcloud.com/45202
Tested-by: jenkins <devops@whamcloud.com>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: John L. Hammond <jhammond@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Mikhail Pershin [Thu, 14 Oct 2021 14:16:21 +0000 (17:16 +0300)]
LU-15142 lctl: fixes for set_param -P and llog_print
- properly handle permanent param deletion
- don't print skipped parameters in llog_print output
- add --raw option to llog_print to output all entries
including markers
Signed-off-by: Mikhail Pershin <mpershin@whamcloud.com>
Change-Id: Id93a206a255dc885343efa293e1ee2672493e5e5
Reviewed-on: https://review.whamcloud.com/45332
Reviewed-by: John L. Hammond <jhammond@whamcloud.com>
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Patrick Farrell [Sun, 7 Nov 2021 22:04:40 +0000 (17:04 -0500)]
LU-15197 llite: Do not count tiny write twice
We accidentally count bytes written with tiny write twice
in stats. Remove the extra count.
This also has the positive effect of improving tiny write
performance by about 4% by removing an extra call to the
stats code (the main cost is ktime_get()).
Before, 8 byte dd:
13.9 MiB/s
After:
14.3 MiB/s
Test-parameters: trivial
Signed-off-by: Patrick Farrell <pfarrell@whamcloud.com>
Change-Id: Ia11e7f16e3e3d0c4012f87cde817ad7b21128fa8
Reviewed-on: https://review.whamcloud.com/45476
Tested-by: jenkins <devops@whamcloud.com>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Aurelien Degremont <degremoa@amazon.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Chris Horn [Fri, 22 Oct 2021 01:51:40 +0000 (01:51 +0000)]
LU-15152 tests: auster reports wrong testsuite status
auster always reports testsuites returned 0 even when there
are failures.
Test-Parameters: trivial testlist=sanity-lnet env=ONLY=230,ONLY_REPEAT=20
Signed-off-by: Chris Horn <chris.horn@hpe.com>
Change-Id: I7a8101d6cfd854d8419edf55c18a72e211f5e5c8
Reviewed-on: https://review.whamcloud.com/45343
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: James Nunez <jnunez@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Serguei Smirnov [Wed, 23 Jun 2021 22:51:21 +0000 (15:51 -0700)]
LU-14662 lnet: set eth routes needed for multi rail
When ksocklnd is initialized or new ethernet interfaces
are added via lnetctl, set the routing rules using a common
shell script ksocklnd-config. This ensures control over
source interface when sending traffic.
For example, for eth0 with ip 192.168.122.142/24:
the output of "ip route show table eth0" should be
192.168.122.0/24 dev eth0 proto kernel scope link src 192.168.122.142
This step can be omitted by specifying
options ksocklnd skip_mr_route_setup=1
in the conf file, or by using switch
--skip-mr-route-setup
when adding NI with lnetctl. Note that the module parameter
takes priority over the lnetctl switch: if skip-mr-route-setup
is not specified when adding NI with lnetctl, the route still
won't get created if the conf file has skip_mr_route_setup=1.
The route also won't be created if any route already exists
for the given interface, assuming advanced users who manage
routing on their own will want to continue doing so.
Test-Parameters: trivial
Signed-off-by: Serguei Smirnov <ssmirnov@whamcloud.com>
Change-Id: Ia14e637bd29d4bbce5dd93daad9992336b2e6b15
Reviewed-on: https://review.whamcloud.com/44065
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Amir Shehata <ashehata@whamcloud.com>
Reviewed-by: Cyril Bordage <cbordage@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Lai Siyao [Sat, 23 Oct 2021 14:44:36 +0000 (10:44 -0400)]
LU-14448 lod: verify LOV early in lod_get_default_striping
lod_get_default_striping() will get both default LOV and default LMV,
and parse them to struct lod_default_striping one by one, however the
LOV and LMV data are both stored in lod_thread_info.lti_ea_store, so
lod_verify_striping() should verify LOV upon getting LOV, otherwise
if both exists, it's LMV that's verified, which will return -EINVAL.
Fixes:
6a08df2d0effc7a ("LU-14448 lod: verify LOV before set/inherit")
Signed-off-by: Lai Siyao <lai.siyao@whamcloud.com>
Change-Id: I9763d35bdbc74101fa8515d5096ec457a4cb3524
Reviewed-on: https://review.whamcloud.com/45370
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Tested-by: jenkins <devops@whamcloud.com>
Reviewed-by: Hongchao Zhang <hongchao@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Vladimir Saveliev [Wed, 3 Nov 2021 10:52:14 +0000 (13:52 +0300)]
LU-9704 grant: ignore grant info on read resend
The following scenario makes a message like "claims 28672 GRANT, real
grant 0" to appear:
1. client owns X grants and run rpcs to shrink part of those
2. server fails over so that the shrink rpc is to be resent.
3. on the clinet reconnect server and client sync on initial amount
of grants for the client.
4. shrink rpc is resend, if server disk space is enough, shrink does
not happen and the client adds amount of grants it was going to
shrink to its newly initial amount of grants. Now, client thinks that
it owns more grants than it does from server points of view.
5. the client consumes grants and sends rpcs to server. Server avoids
allocating new grants for the client if the current amount of grant
is big enough:
static long tgt_grant_alloc(struct obd_export *exp, u64 curgrant,
...
if (curgrant >= want || curgrant >= ted->ted_grant + chunk)
RETURN(0);
6. client continues grants consuming which eventually leads to
complains like "claims 28672 GRANT, real grant 0".
In case of resent of read and set_info:shrink RPCs grant info should
be ignored as it was reset on reconnect.
Tests to illustrate the issue is added.
HPE-bug-id: LUS-7666
Change-Id: I8af1db287dc61c713e5439f4cf6bd652ce02c12c
Signed-off-by: Vladimir Saveliev <vlaidimir.saveliev@hpe.com>
Reviewed-on: https://review.whamcloud.com/45371
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Mike Pershin <mpershin@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Andreas Dilger [Mon, 19 Apr 2021 21:32:31 +0000 (15:32 -0600)]
LU-930 doc: update lfs migrate usage and man page
Update the usage and man page for "lfs migrate -m", noting that
this command will recursively migrate an entire directory tree.
It is not currently possible to migrate files with DoM components
between MDTs, so provide an example of how to work around this.
Only print the command-line options for commands in the usage
message instead of the full usage, since it is otherwise much
too verbose to see the actual error message being printed. The
user should read the lfs-migrate.1 man page for full usage.
Test-Parameters: trivial
Signed-off-by: Andreas Dilger <adilger@whamcloud.com>
Change-Id: I9e34a33fcc3f0e2b90bc499fb7b946c53e6111d1
Reviewed-on: https://review.whamcloud.com/43378
Tested-by: jenkins <devops@whamcloud.com>
Reviewed-by: James Nunez <jnunez@whamcloud.com>
Reviewed-by: Lai Siyao <lai.siyao@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Andriy Skulysh [Fri, 30 Aug 2019 11:43:29 +0000 (14:43 +0300)]
LU-15121 llite: skip request slot for lmv_revalidate_slaves()
Some syscalls need lmv_revalidate_slaves(). It requires
second lock enqueue and the it can be blocked by
lack of RPC slots.
Don't acquire rpc slot for second lock enqueue.
Change-Id: Ida23c648c2bd169c4d238543731796232aa490dc
HPE-bug-id: LUS-8416
Signed-off-by: Andriy Skulysh <c17819@cray.com>
Reviewed-by: Vitaly Fertman <c17818@cray.com>
Reviewed-by: Alexander Zarochentsev <c17826@cray.com>
Reviewed-on: https://review.whamcloud.com/45275
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Vitaly Fertman <vitaly.fertman@hpe.com>
Reviewed-by: Alexander Zarochentsev <alexander.zarochentsev@hpe.com>
Reviewed-by: Lai Siyao <lai.siyao@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Sebastien Buisson [Thu, 21 Oct 2021 06:56:44 +0000 (08:56 +0200)]
LU-15141 quota: optimize capability check for root squash
On client side, checking for owner/group quota can be directly
bypassed if this is for root and there is no root squash.
Change-Id: If29eca428d8748df412a717615e4d0a4886ddd04
Fixes:
a4fbe7341b ("LU-14739 quota: nodemap squashed root cannot bypass quota")
Signed-off-by: Sebastien Buisson <sbuisson@ddn.com>
Reviewed-on: https://review.whamcloud.com/45322
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-by: Hongchao Zhang <hongchao@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Andreas Dilger [Sat, 30 Oct 2021 00:40:40 +0000 (18:40 -0600)]
LU-12678 ptlrpc: remove bogus LASSERT
In the error case, it isn't possible for rc to be both -ENOMEM and
0 at the same time, so remove the incorrect LASSERT(rc == 0) to
avoid crashing the system on an allocation failure.
Improve error messages to conform to code style.
Fixes:
ceeeae4271fd ("LU-12678 lnet: me: discard struct lnet_handle_me")
Signed-off-by: Andreas Dilger <adilger@whamcloud.com>
Change-Id: I61ac5d735d7b2658dae76213a2d40cbfd2bb8bb9
Reviewed-on: https://review.whamcloud.com/45421
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: James Simmons <jsimmons@infradead.org>
Reviewed-by: Chris Horn <chris.horn@hpe.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Xinliang Liu [Thu, 28 Oct 2021 09:48:38 +0000 (09:48 +0000)]
LU-11667 tests: Fix sanity test 317 for 64K PAGE_SIZE OST
When create a file, blocks are allocated with PAGE_SIZE aligned,
see function osd_ldiskfs_map_inode_pages(). E.g. for 64K PAGE_SIZE
Arm64 OST server, if create a file with size less than 64K, it
actually allocates 128 blocks each block 512 Bytes.
It needs to adjust the test for 64K PAGE_SIZE OST server.
Test-Parameters: trivial
Test-Parameters: clientarch=aarch64 fstype=ldiskfs testlist=sanity \
env=PTLDEBUG=-1,ONLY=317
Test-Parameters: fstype=ldiskfs testlist=sanity \
env=PTLDEBUG=-1,ONLY=317
Change-Id: Iada701f4f424093e847fc70aa843873b75fe6b06
Signed-off-by: Xinliang Liu <xinliang.liu@linaro.org>
Reviewed-on: https://review.whamcloud.com/45395
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-by: Arshad Hussain <arshad.hussain@aeoncomputing.com>
Tested-by: jenkins <devops@whamcloud.com>
Reviewed-by: James Simmons <jsimmons@infradead.org>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Yang Sheng [Tue, 26 Oct 2021 08:09:20 +0000 (16:09 +0800)]
LU-15156 kernel: back port patch for rwsem issue
RHEL7 included a defeat in rwsem. It can cause a
thread hung on rwsem waiting infinity. Backport
commit:
5c1ec49b60cdb31e51010f8a647f3189b774bddf
to fix this issue.
Signed-off-by: Yang Sheng <ys@whamcloud.com>
Change-Id: Ic5c469ce744ad5882c13163a9bfe14faef8fd446
Reviewed-on: https://review.whamcloud.com/45383
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-by: Alex Zhuravlev <bzzz@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Tested-by: jenkins <devops@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Alex Zhuravlev [Wed, 27 Oct 2021 05:48:03 +0000 (08:48 +0300)]
LU-15168 osd: use large allocation for idc cache
as in some cases (e.g. ofd precreate) the cache can grow to dozens
of kilobytes (sizeof(struct idc_map_cache)=40 * 1024).
Signed-off-by: Alex Zhuravlev <bzzz@whamcloud.com>
Change-Id: Id9e0996a7a1d07065f4a50c1d5be5051e756559a
Reviewed-on: https://review.whamcloud.com/45382
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Mike Pershin <mpershin@whamcloud.com>
Tested-by: jenkins <devops@whamcloud.com>
Jian Yu [Mon, 25 Oct 2021 17:33:03 +0000 (10:33 -0700)]
LU-15154 kernel: kernel update SLES15 SP3 [5.3.18-59.27.1]
Update SLES15 SP3 kernel to 5.3.18-59.27.1 for Lustre client.
Test-Parameters: trivial
Change-Id: Ie3c369a8e93a75b4afbde55489bd3819bb39e1de
Signed-off-by: Jian Yu <yujian@whamcloud.com>
Reviewed-on: https://review.whamcloud.com/45349
Reviewed-by: Minh Diep <mdiep@whamcloud.com>
Reviewed-by: Yang Sheng <ys@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Tested-by: jenkins <devops@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Chris Horn [Fri, 22 Oct 2021 00:13:19 +0000 (00:13 +0000)]
LU-15138 lnet: Fail peer add for existing gw peer
If there's an existing peer entry for a peer that is being added
via CLI, and that existing peer was not created via the CLI, then
DLC will attempt to delete the existing peer before creating a new
one. The exit status of the peer deletion was not being checked.
This results in the ability to add duplicate peers for gateways,
because gateways cannot be deleted via lnetctl unless the routes
for that gateway have been removed first.
Test-Parameters: trivial testlist=sanity-lnet
Signed-off-by: Chris Horn <chris.horn@hpe.com>
Change-Id: I9b7864a2bd21540336f72d96e180c89bd0aae8dc
Reviewed-on: https://review.whamcloud.com/45337
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Alexander Boyko <alexander.boyko@hpe.com>
Reviewed-by: Andriy Skulysh <andriy.skulysh@hpe.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Alex Zhuravlev [Wed, 20 Oct 2021 11:10:57 +0000 (14:10 +0300)]
LU-15086 ptlrpc: fix timeout after spurious wakeup
so that final timeout don't exceed requested one
Signed-off-by: Alex Zhuravlev <bzzz@whamcloud.com>
Change-Id: Iff5e08c589cbbc3c483915002f3f9df7a6f2678a
Reviewed-on: https://review.whamcloud.com/45308
Reviewed-by: Andriy Skulysh <andriy.skulysh@hpe.com>
Reviewed-by: James Simmons <jsimmons@infradead.org>
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Xinliang Liu [Tue, 19 Oct 2021 08:15:59 +0000 (08:15 +0000)]
LU-15122 osd-ldiskfs: Fix ASSERTION( iobuf->dr_rw == 0 ) with 64KB PAGE_SIZE
During a writing, if there is a page can not be mapped to blocks
at once, it will cause "ASSERTION( iobuf->dr_rw == 0 )" crash
which leads by the overflow access of mapped blocks array.
This will happen on Arm platforms easily with 64KB PAGE_SIZE.
And will not happen on x86_64 platforms with 4KB PAGE_SIZE.
Because for 4KB block size, if PAGE_SIZE is 4KB, then i == 0
and blocks_left_page == 1. Which makes the inner loop each time
handle one block. Thus the outer loop condition "block_idx <
block_idx_end;" insures blocks[] array access not overflow.
Check the actual mapped count so that mapped blocks array
access will not overflow.
Fixes:
0271b17b80a8 ("LU-14134 osd-ldiskfs: reduce credits for new writing")
Change-Id: Icd46c04bea2d7930456840694d422758eebb4186
Signed-off-by: Xinliang Liu <xinliang.liu@linaro.org>
Reviewed-on: https://review.whamcloud.com/45288
Tested-by: jenkins <devops@whamcloud.com>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-by: James Simmons <jsimmons@infradead.org>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Li Dongyang <dongyangli@ddn.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Chris Horn [Wed, 13 Oct 2021 23:30:01 +0000 (18:30 -0500)]
LU-15102 lnet: Reset ni_ping_count only on receive
The lnet_ni:ni_ping_count is currently reset on every (healthy) tx.
We should only reset it when receiving a message over the NI. Taking
net_lock 0 on every tx results in a performance loss for certain
workloads.
Test-Parameters: trivial testlist=sanity-lnet
Fixes:
8fdf2bc62a ("LU-13569 lnet: Recover local NI w/exponential backoff interval")
HPE-bug-id: LUS-10427
Signed-off-by: Chris Horn <chris.horn@hpe.com>
Change-Id: I67ea3aa977cb5d67b04f7957120c29e9985c83e6
Reviewed-on: https://review.whamcloud.com/45235
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Serguei Smirnov <ssmirnov@whamcloud.com>
Reviewed-by: Andriy Skulysh <andriy.skulysh@hpe.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Chris Horn [Thu, 16 Sep 2021 17:12:38 +0000 (12:12 -0500)]
LU-15092 o2iblnd: Fix logic for unaligned transfer
It's possible for there to be an offset for the first page of a
transfer. However, there are two bugs with this code in o2iblnd.
The first is that this use-case will require LNET_MAX_IOV + 1 local
RDMA fragments, but we do not specify the correct corresponding values
for the max page list to ib_alloc_fast_reg_page_list(),
ib_alloc_fast_reg_mr(), etc.
The second issue is that the logic in kiblnd_setup_rd_kiov() attempts
to obtain one more scatterlist entry than is actually needed. This
causes the transfer to fail with -EFAULT.
Test-Parameters: trivial
HPE-bug-id: LUS-10407
Fixes:
d226464aca ("LU-8057 ko2iblnd: Replace sg++ with sg = sg_next(sg)")
Signed-off-by: Chris Horn <chris.horn@hpe.com>
Change-Id: Ifb843f11ae34a99b7d8f93d94966e3dfa1ce90e5
Reviewed-on: https://review.whamcloud.com/45216
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: James Simmons <jsimmons@infradead.org>
Reviewed-by: Andriy Skulysh <andriy.skulysh@hpe.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Chris Horn [Wed, 29 Sep 2021 17:42:26 +0000 (12:42 -0500)]
LU-15094 o2iblnd: map_on_demand not needed for frag interop
The map_on_demand tunable is not used for setting max frags so don't
require that it be set in order to negotiate max frags.
HPE-bug-id: LUS-10488
Test-Parameters: trivial
Signed-off-by: Chris Horn <chris.horn@hpe.com>
Change-Id: Ie89f1f035f4b05244feffb848c14582a8c7cf0e6
Reviewed-on: https://review.whamcloud.com/45215
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Serguei Smirnov <ssmirnov@whamcloud.com>
Reviewed-by: Andriy Skulysh <andriy.skulysh@hpe.com>
Reviewed-by: Alexey Lyashkov <alexey.lyashkov@hpe.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Sergey Cheremencev [Thu, 17 Jun 2021 10:45:42 +0000 (13:45 +0300)]
LU-15049 quota: fix a panic with pool number > 16
Fix a panic that may occur when there are more than 16
pools in a system:
qti_pools_add()) ASSERTION( qti->qti_pools_num >= QMT_MAX_POOL_NUM ) failed: Forgot init?
ffff91a5f9625800
HPE-bug-id: LUS-10116
Change-Id: I4f73b74d2fd3e85a51cf3c30e2eec29645f164be
Reviewed-by: Petros Koutoupis <pkoutoupis@cray.com>
Reviewed-by: Shaun Tancheff <stancheff@cray.com>
Tested-by: Elena Gryaznova <c17455@cray.com>
Signed-off-by: Sergey Cheremencev <sergey.cheremencev@hpe.com>
Reviewed-on: https://review.whamcloud.com/45105
Reviewed-by: Petros Koutoupis <petros.koutoupis@hpe.com>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Sergey Cheremencev [Thu, 22 Jul 2021 10:56:24 +0000 (13:56 +0300)]
LU-15048 quota: check that qti_lqes has been inited
qti_lqes_resotre_init/fini should check that qti_lqes
has been inited before address qti_lqes_count.
Fix helps against following panic:
qti_lqes_restore_fini()) ASSERTION( qmt_info(env)->qti_lqes_rstr ) failed:
HPE-bug-id: LUS-10239
Change-Id: Ic93d87535f615fe419b2c3a2453506c515837031
Reviewed-on: https://es-gerrit.dev.cray.com/159116
Reviewed-by: Alexander Boyko <alexander.boyko@hpe.com>
Reviewed-by: Shaun Tancheff <stancheff@cray.com>
Tested-by: Elena Gryaznova <c17455@cray.com>
Signed-off-by: Sergey Cheremencev <sergey.cheremencev@hpe.com>
Reviewed-on: https://review.whamcloud.com/45102
Tested-by: Maloo <maloo@whamcloud.com>
Tested-by: jenkins <devops@whamcloud.com>
Reviewed-by: Shaun Tancheff <shaun.tancheff@hpe.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Bobi Jam [Thu, 2 Sep 2021 16:38:43 +0000 (00:38 +0800)]
LU-14713 llite: tighten condition for fault not drop mmap_sem
As __lock_page_or_retry() indicates, filemap_fault() will return
VM_FAULT_RETRY without releasing mmap_sem iff flags contains
FAULT_FLAG_ALLOW_RETRY and FAULT_FLAG_RETRY_NOWAIT.
So ll_fault0() should pass in FAULT_FLAG_ALLOW_RETRY |
FAULT_FLAG_RETRY_NOWAIT in ll_filemap_fault() so that when it
returns VM_FAULT_RETRY, we can pass on trying normal fault
under DLM lock as mmap_sem is still being held.
While in Linux 5.1 (
6b4c9f4469819) FAULT_FLAG_RETRY_NOWAIT is enough
to not drop mmap_sem when failed to lock the page.
Fixes:
87865e4ae9 ("LU-13182 llite: Avoid eternel retry loops with MAP_POPULATE")
Signed-off-by: Bobi Jam <bobijam@whamcloud.com>
Change-Id: I9420c587301722b597155558657577349a8141e4
Reviewed-on: https://review.whamcloud.com/44715
Reviewed-by: Patrick Farrell <pfarrell@whamcloud.com>
Tested-by: jenkins <devops@whamcloud.com>
Reviewed-by: Neil Brown <neilb@suse.de>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
James Simmons [Mon, 25 Oct 2021 20:56:31 +0000 (16:56 -0400)]
LU-14927 osd: share brw_stats code between OSD back ends.
Both the ldiskfs and ZFS OSD backend handle brw_stats. With the
stricter GPL requirement ZFS can no longer carry the brw_stats
code. So move the common code to lprocfs_status_server.c as
well as move brw_stats to debugfs as well.
Change-Id: I294e5df3557552266dd3a02d3bc9844c42c01f60
Signed-off-by: James Simmons <jsimmons@infradead.org>
Reviewed-on: https://review.whamcloud.com/44690
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-by: Alex Zhuravlev <bzzz@whamcloud.com>
Reviewed-by: Aurelien Degremont <degremoa@amazon.com>
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Qian Yingjin [Fri, 25 Jun 2021 08:22:35 +0000 (16:22 +0800)]
LU-14793 hsm: record index for further HSM action scanning
there is contention between HSM archive request and "hsm_cdtr"
kernel thread:
->mdt_hsm_request()
->mdt_hsm_add_actions()
->mdt_hsm_register_hal()
->mdt_agent_record_add()
->down_write(&cdt->cdt_llog_lock)
->llog_cat_add()
->up_write(&cdt->cdt_llog_lock)
->mdt_coordinator()
->cdt_llog_process()
->down_write(&cdt->cdt_llog_lock);
->llog_cat_process()
->up_write(&cdt->cdt_llog_lock);
HSM archive request and HSM cat llog scanning in the kernel daemon
"hsm_cdtr" are both contenting for write llog lock to add or
update the "hsm_actions" llog.
In the tesing, it uses max_requests = 1000000.
In the current implementation, it means kernel daemon thread
"hsm_cdtr" needs to scan nearly whole "hsm_actions" llog from the
beginning position with write llog lock held.
This will slow down the HSM archive requests which is contented
for write llog lock.
As llog is append-only, we record the latest handled position in
the llog, thus next scanning can start from the previous recorded
postion (llog index), does not need to start from the beginning.
Another way to mitigate this probelm is:
when the llog scanner found that there are other process
contended for the llog lock, it will stop the llog scanning and
release write llog lock properly for incoming HSM archive requests.
After applied this patch, with 200000 HSM actions in llog, the time
to queue 10000 HSM archive requests reduces from 10 seconds to 4
seconds.
Signed-off-by: Qian Yingjin <qian@ddn.com>
Change-Id: I2e92daf34844605ee648787daf859143335c68bf
Reviewed-on: https://review.whamcloud.com/44077
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-by: Sergey Cheremencev <sergey.cheremencev@hpe.com>
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Shaun Tancheff [Fri, 30 Apr 2021 11:43:46 +0000 (06:43 -0500)]
LU-14706 libcfs: ___wait_event() update for older kernels
A couple of wait issues fail to build correctly on older
3.0 kernels
Fixes:
f6df31a163 ("LU-10467 lustre: add wait_event macros suitable for upstream")
Fixes:
d05427a785 ("LU-10428 lnet: call event handlers without res_lock")
HPE-bug-id: LUS-9975
Signed-off-by: Shaun Tancheff <shaun.tancheff@hpe.com>
Change-Id: I275fbf79c4472575d867075a3e3ebd3d6ec1cdfa
Reviewed-on: https://review.whamcloud.com/43785
Reviewed-by: Neil Brown <neilb@suse.de>
Reviewed-by: James Simmons <jsimmons@infradead.org>
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Alex Zhuravlev [Tue, 30 Mar 2021 05:57:14 +0000 (08:57 +0300)]
LU-13358 libcfs: add timeout to cfs_race() to fix race
there is no guarantee for the branches in cfs_race() to be executed
in strict order, thus it's possible that the second branch (with
cfs_race_state=1) is executed before the first branch and then another
thread executing the first branch gets stuck.
this construction is used for testing only and as a
workaround it's enough to timeout.
Signed-off-by: Alex Zhuravlev <bzzz@whamcloud.com>
Change-Id: Ie1cc0accedb3e1a198d4b17d1ab00ce298c560f2
Reviewed-on: https://review.whamcloud.com/43161
Reviewed-by: James Simmons <jsimmons@infradead.org>
Reviewed-by: Neil Brown <neilb@suse.de>
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Andrew Perepechko [Sun, 31 Oct 2021 20:03:30 +0000 (23:03 +0300)]
LU-15171 osd-ldiskfs: xattr_sem locking is missing for dquot_transfer
Kernel commit
7a9ca53ae (~v4.13) added the requirement for xattr_sem locking
when calling *dquot_transfer. As of now, in rare cases, it is possible that
we can modify inode xattrs and perform their consistency checks in parallel,
which can fail.
Change-Id: I041694e30ce6c8398864c0ad57671df0bffd2f52
Signed-off-by: Andrew Perepechko <andrew.perepechko@hpe.com>
HPE-bug-id: LUS-10549
Reviewed-on: https://review.whamcloud.com/45424
Reviewed-by: Alexander Zarochentsev <alexander.zarochentsev@hpe.com>
Tested-by: jenkins <devops@whamcloud.com>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Andrew Perepechko [Thu, 21 Oct 2021 19:40:39 +0000 (22:40 +0300)]
LU-15143 osd-ldiskfs: osd_declare_write() underestimates credits
osd_declare_write() seems to underestimate journal credits for
the extent case. It does not properly account credits for
a new extent tree block. 3 credits (bitmap, gd, self) should be
accounted for a new data block and for a new extent tree block.
Change-Id: Iad463cac3a2a8c2b2a6b1a634e7502039bb1b7e5
HPE-bug-id: LUS-10514
Signed-off-by: Andrew Perepechko <andrew.perepechko@hpe.com>
Reviewed-on: https://review.whamcloud.com/45328
Reviewed-by: Alexander Boyko <alexander.boyko@hpe.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Artem Blagodarenko <artem.blagodarenko@hpe.com>
Tested-by: jenkins <devops@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Mr NeilBrown [Fri, 17 Apr 2020 01:27:54 +0000 (11:27 +1000)]
LU-10391 lnet: switch to large lnet_processid for matching
Change lnet_libhandle.me_match_id and lnet_match_info.mi_id to
struct lnet_processid, so they support large nids.
This requires changing
LNetMEAttach(), lnet_mt_match_head(), lnet_mt_of_attach(),
lnet_ptl_match_type(), lnet_match2mt()
to accept a pointer to lnet_processid rather than an
lnet_process_id.
Test-Parameters: trivial
Test-Parameters: serverversion=2.12 serverdistro=el7.9 testlist=runtests
Test-Parameters: clientversion=2.12 testlist=runtests
Signed-off-by: Mr NeilBrown <neilb@suse.de>
Change-Id: I6957b467bb9af84e20a4525db6351694f4d2a7af
Reviewed-on: https://review.whamcloud.com/43597
Reviewed-by: James Simmons <jsimmons@infradead.org>
Reviewed-by: Chris Horn <chris.horn@hpe.com>
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Amir Shehata <ashehata@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Mr NeilBrown [Sat, 11 Sep 2021 14:16:49 +0000 (10:16 -0400)]
LU-10391 lnet: extend prefered nids in struct lnet_peer_ni
union lpni_pref in struct lnet_peer_ni how has 'struct lnet_nid'
rather than lnet_nid_t.
Also, lnet_peer_ni_set_no_mr_pref_nid() allows the pref nid to be NULL
and is a no-op in that case.
Rather then updating the user-space cfs_match_nid_net() in
libcfs/utils/nidstrings.c, remove it as it is unused.
Test-Parameters: trivial
Test-Parameters: serverversion=2.12 serverdistro=el7.9 testlist=runtests
Test-Parameters: clientversion=2.12 testlist=runtests
Signed-off-by: Mr NeilBrown <neilb@suse.de>
Change-Id: I9a2453185aa5d708e6939dadc1e954c9dbd24efc
Reviewed-on: https://review.whamcloud.com/43596
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: James Simmons <jsimmons@infradead.org>
Reviewed-by: Chris Horn <chris.horn@hpe.com>
Reviewed-by: Amir Shehata <ashehata@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Mr NeilBrown [Mon, 6 Apr 2020 03:11:33 +0000 (13:11 +1000)]
LU-10391 lnet: change tp_nid to 16byte in lnet_test_peer.
This updates 'struct lnet_test_peer' to store a large address nid.
Test-Parameters: trivial
Test-Parameters: serverversion=2.12 serverdistro=el7.9 testlist=runtests
Test-Parameters: clientversion=2.12 testlist=runtests
Signed-off-by: Mr NeilBrown <neilb@suse.de>
Change-Id: Id2f97a841bb0738503b0e87e7a9e2f8bebc4c3ec
Reviewed-on: https://review.whamcloud.com/43595
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: James Simmons <jsimmons@infradead.org>
Reviewed-by: Chris Horn <chris.horn@hpe.com>
Reviewed-by: Amir Shehata <ashehata@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Serguei Smirnov [Fri, 12 Nov 2021 17:08:53 +0000 (09:08 -0800)]
LU-15210 tests: fix sanity-lnet to handle duplicate IP
The same IP may be added on the same interface with different netmasks
specified.
sanity-lnet is not expecting this and fails. Fix it.
Test-parameters: trivial testlist=sanity-lnet
Signed-off-by: Serguei Smirnov <ssmirnov@whamcloud.com>
Change-Id: I70c2df5f14c88362ea5af6f06410823c56535dee
Reviewed-on: https://review.whamcloud.com/45551
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Andriy Skulysh [Mon, 16 Dec 2019 20:09:37 +0000 (22:09 +0200)]
LU-13456 ldlm: fix reprocessing of locks with more bits
Reprocessing check queues should be extended
with just granted lock inodebits.
ldlm_reprocess_all() should be called on BL AST race.
Change-Id: Ifd232062068481c1c62fa2f2a14c7778d4402260
Fixes:
2250e072c3785 ("LU-12017 ldlm: DoM truncate deadlock")
HPE-bug-id: LUS-8224
Signed-off-by: Andriy Skulysh <c17819@cray.com>
Reviewed-by: Alexander Boyko <c17825@cray.com>
Reviewed-by: Vitaly Fertman <c17818@cray.com>
Reviewed-on: https://review.whamcloud.com/38244
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Vitaly Fertman <vitaly.fertman@hpe.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Lai Siyao [Tue, 12 Oct 2021 22:15:37 +0000 (18:15 -0400)]
LU-15070 llite: update default LMV upon any change
max_inherit and max_inherit_rr was newly added, and they are missing
in lsm_md_eq(), therefore client may not update default LMV when
either of these two fields is changed.
Add sanityn 112.
Signed-off-by: Lai Siyao <lai.siyao@whamcloud.com>
Change-Id: Iac71b530b3702105c4213715826b1782c6aba7ca
Reviewed-on: https://review.whamcloud.com/45237
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Hongchao Zhang <hongchao@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Lai Siyao [Tue, 12 Oct 2021 22:20:21 +0000 (18:20 -0400)]
LU-15070 mdt: revoke remote LOOKUP lock for default LMV
When setting default LMV, it will revoke LOOKUP lock, while if dir
is remote dir, its LOOKUP lock is on MDT where its parent is located.
Signed-off-by: Lai Siyao <lai.siyao@whamcloud.com>
Change-Id: I9f079a0bcff530603725ce72cd89c14935ba913b
Reviewed-on: https://review.whamcloud.com/45236
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-by: Hongchao Zhang <hongchao@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Elena Gryaznova [Tue, 26 Oct 2021 11:34:30 +0000 (14:34 +0300)]
LU-15151 tests: use facet check instead of node check
Change wait_update_cond() call to wait_update_facet_cond()
call in test_119.
Fixes:
98f107b53e4d ("LU-9699 osp: don't assert on OSP duplicating")
Test-Parameters: trivial testlist=conf-sanity env=ONLY=119
Signed-off-by: Elena Gryaznova <elena.gryaznova@hpe.com>
HPE-bug-id: LUS-10557
Change-Id: Ia1ff7921026212814ec71e0c3aa60f23fbd7278f
Reviewed-on: https://review.whamcloud.com/45369
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-by: Vladimir Saveliev <vlaidimir.saveliev@hpe.com>
Tested-by: Maloo <maloo@whamcloud.com>
Tested-by: jenkins <devops@whamcloud.com>
Jian Yu [Mon, 25 Oct 2021 20:04:53 +0000 (13:04 -0700)]
LU-15160 kernel: kernel update SLES12 SP5 [4.12.14-122.91.2]
Update SLES12 SP5 kernel to 4.12.14-122.91.2 for Lustre client.
Test-Parameters: trivial clientdistro=sles12sp5 \
env=SANITY_EXCEPT="56oc 430c 817" testlist=sanity
Change-Id: Ia6620869fa84d72f8d22c4a8a039600037ddb2d9
Signed-off-by: Jian Yu <yujian@whamcloud.com>
Reviewed-on: https://review.whamcloud.com/45358
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Minh Diep <mdiep@whamcloud.com>
Reviewed-by: Yang Sheng <ys@whamcloud.com>
Tested-by: jenkins <devops@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Chris Horn [Fri, 22 Oct 2021 01:07:58 +0000 (01:07 +0000)]
LU-15149 lnet: Missing newline in lnet_add_route
CWARN string is missing a newline character.
Test-Parameters: trivial
Fixes:
3f2844dc93 ("LU-14945 lnet: don't use hops to determine the route state")
Signed-off-by: Chris Horn <chris.horn@hpe.com>
Change-Id: I06370c36e9d88b7e02e000bfb573297ff281aef1
Reviewed-on: https://review.whamcloud.com/45340
Reviewed-by: Serguei Smirnov <ssmirnov@whamcloud.com>
Reviewed-by: James Simmons <jsimmons@infradead.org>
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Serguei Smirnov [Thu, 21 Oct 2021 02:09:06 +0000 (19:09 -0700)]
LU-15136 socklnd: default conns_per_peer to 0
Setting conns_per_peer to 0 triggers socklnd to choose the
(heuristically) optimal setting for the interface given its speed.
Make 0 the default for socklnd conns_per_peer.
Test-parameters: trivial testlist=sanity-lnet
Signed-off-by: Serguei Smirnov <ssmirnov@whamcloud.com>
Fixes:
c44afcfb72 ("LU-12815 socklnd: set conns_per_peer based on link speed")
Change-Id: Ie6e76eaee8693472384cce362b394b216142884e
Reviewed-on: https://review.whamcloud.com/45319
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-by: Chris Horn <chris.horn@hpe.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: James Simmons <jsimmons@infradead.org>
Tested-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Chris Horn [Fri, 22 Oct 2021 01:34:23 +0000 (01:34 +0000)]
LU-15150 tests: sanity-lnet removes testsuite log on failure
cleanup_testsuite() needs to be more selective when removing files
created by sub-tests.
Test-Parameters: trivial testlist=sanity-lnet
Fixes:
aa739144551 ("LU-13569 tests: Check LNet Health recovery logic")
Signed-off-by: Chris Horn <chris.horn@hpe.com>
Change-Id: Ic17a68ff2aa552594a0f1ea470c39177abe985fc
Reviewed-on: https://review.whamcloud.com/45342
Reviewed-by: Serguei Smirnov <ssmirnov@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Tested-by: jenkins <devops@whamcloud.com>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Elena Gryaznova [Tue, 19 Oct 2021 18:10:31 +0000 (21:10 +0300)]
LU-15131 tests: check parameter correctly
sanity test_254() always skipped due to wrong
parameter path: MDT0 not initialized before get_param.
Fixes:
33aad7829de5 ("LU-10461 tests: call exit in the skip routine")
Test-Parameters: trivial testlist=sanity
Signed-off-by: Elena Gryaznova <elena.gryaznova@hpe.com>
HPE-bug-id: LUS-9969
Reviewed-by: Alexander Boyko <alexander.boyko@hpe.com>
Reviewed-by: Alexander Zarochentsev <alexander.zarochentsev@hpe.com>
Change-Id: I02e38a27ca4656128e62339d63425f85386fa905
Reviewed-on: https://review.whamcloud.com/45300
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Elena Gryaznova [Tue, 19 Oct 2021 17:36:06 +0000 (20:36 +0300)]
LU-15061 tests: fix sanity-dom exit status
If any of sanity tests fails during "ONLY=sanity sh sanity-dom.sh"
execution -- the next "ONLY=sanityn sh sanity-dom.sh" is always
detected as failed even if all sanityn tests pass. This is
because of ${TMP}/sanity.log exists and is checked for
"FAIL" also, while ${TMP}/sanityn.log is to be proceeded only.
Fixes
e2cb43c409b9 ("LU-13773 tests: subscript failure propagation")
Test-Parameters: trivial testlist=sanity-dom
HPE-bug-id: LUS-10480
Signed-off-by: Elena Gryaznova <elena.gryaznova@hpe.com>
Change-Id: I764d1d6df08da13acfecd445e67ced1b455ddce8
Reviewed-on: https://review.whamcloud.com/45299
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Jian Yu <yujian@whamcloud.com>
Reviewed-by: Alexander Boyko <alexander.boyko@hpe.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Sebastien Buisson [Tue, 19 Oct 2021 15:59:33 +0000 (17:59 +0200)]
LU-15098 tests: sanity-sec 27a exec commands on right node
In nodemap_exercise_fileset called from sanity-sec test 27a,
make sure all commands are executed on first client, as we are
testing properties of nodemaps 'default' and 'c0'.
And make sure 'default' nodemap has admin and trusted properties
set to 1, as we are carrying operations as root.
Test-Parameters: trivial
Test-Parameters: testlist=sanity-sec clientcount=2 env=ONLY=27a
Fixes:
0daeebcbdc ("LU-14797 nodemap: map project id")
Signed-off-by: Sebastien Buisson <sbuisson@ddn.com>
Change-Id: Idd9f391db60475721f3a3856b5e3bee1a18bbbca
Reviewed-on: https://review.whamcloud.com/45293
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Jian Yu <yujian@whamcloud.com>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
gaurav mahajan [Thu, 14 Oct 2021 10:12:38 +0000 (13:12 +0300)]
LU-15103 tests: clean up busy cifs mount
Busy cifs mount point makes cleanup_cifs fail which
will infact fails lustre unmount as cifs mount point is
lustre mount.
Test-Parameters: trivial
Signed-off-by: gaurav mahajan <gaurav.mahajan@seagate.com>
Signed-off-by: Elena Gryaznova <elena.gryaznova@hpe.com>
HPE-bug-id: LUS-4146
Reviewed-by: Sergey Cheremencev <sergey.cheremencev@hpe.com>
Reviewed-by: Alexander Zarochentsev <alexander.zarochentsev@hpe.com>
Change-Id: I4b7ec7e0a6a706198e328dad337648bf3cb2c3be
Reviewed-on: https://review.whamcloud.com/45238
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Jian Yu [Wed, 13 Oct 2021 16:43:11 +0000 (09:43 -0700)]
LU-15099 kernel: kernel update RHEL7.9 [3.10.0-1160.45.1.el7]
Update RHEL7.9 kernel to 3.10.0-1160.45.1.el7.
Test-Parameters: trivial clientdistro=el7.9 serverdistro=el7.9
Change-Id: I11c307bfd6a6b353bc7b6fe40bb5d604bc9b3fdc
Signed-off-by: Jian Yu <yujian@whamcloud.com>
Reviewed-on: https://review.whamcloud.com/45226
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Minh Diep <mdiep@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Sebastien Buisson [Wed, 13 Oct 2021 09:35:01 +0000 (09:35 +0000)]
LU-13717 sec: missing defs in includes for client encryption
Add a few missing definitions in lustre_crypto.h that are required
in case Lustre client encryption is built against the in-kernel
fscrypt library.
Fixes:
028281ae19 ("LU-13717 sec: rework includes for client encryption")
Test-Parameters: trivial
Signed-off-by: Sebastien Buisson <sbuisson@ddn.com>
Change-Id: I1965503554dcf660770d201444cfafd54aa84dce
Reviewed-on: https://review.whamcloud.com/45221
Tested-by: jenkins <devops@whamcloud.com>
Reviewed-by: James Simmons <jsimmons@infradead.org>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Chris Horn [Mon, 27 Sep 2021 20:48:02 +0000 (15:48 -0500)]
LU-15093 libcfs: Check if param_set_uint_minmax is provided
Linux kernel v5.15 commit
2a14c9ae15a38148484a128b84bff7e9ffd90d68
moved param_set_uint_minmax to common code.
HPE-bug-id: LUS-10469
Test-Parameters: trivial
Signed-off-by: Chris Horn <chris.horn@hpe.com>
Change-Id: Ifd1d72ae531f0f6c7cd96cc28fbc07c8a8b70886
Reviewed-on: https://review.whamcloud.com/45214
Reviewed-by: Jian Yu <yujian@whamcloud.com>
Reviewed-by: James Simmons <jsimmons@infradead.org>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Petros Koutoupis <petros.koutoupis@hpe.com>
Tested-by: jenkins <devops@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Andriy Skulysh [Sun, 6 Sep 2020 09:21:14 +0000 (12:21 +0300)]
LU-14999 mdt: Deadlock on parent during resend
Parent-child lock order gets broken during resend as
there is child lock already but there isn't parent lock
and MDS tries to lock it again.
Don't lookup child on resend, extract fid from the lock instead.
Change-Id: I443648a8162e770c63fd087dd534c27e7c637c40
HPE-bug-id: LUS-9306
Signed-off-by: Andriy Skulysh <c17819@cray.com>
Reviewed-by: Alexander Boyko <c17825@cray.com>
Reviewed-by: Alexander Zarochentsev <c17826@cray.com>
Reviewed-on: https://review.whamcloud.com/44885
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Alexander Boyko <alexander.boyko@hpe.com>
Reviewed-by: Vitaly Fertman <vitaly.fertman@hpe.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Sebastien Buisson [Wed, 11 Aug 2021 15:44:08 +0000 (17:44 +0200)]
LU-14929 gss: detect libkeyutils dependency
When building GSS support, gss_keyring requires libkeyutils.
So make sure this dependency is properly detected at configure time,
and include keyutils.h only when required.
Signed-off-by: Sebastien Buisson <sbuisson@ddn.com>
Change-Id: I9fa5750f4609250ecdc1c47f68b97bff9be13ace
Reviewed-on: https://review.whamcloud.com/44597
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Shaun Tancheff <shaun.tancheff@hpe.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Vladimir Saveliev [Wed, 20 Oct 2021 10:32:11 +0000 (13:32 +0300)]
LU-14124 target: set OBD_MD_FLGRANT in read's reply
If tgt_grant_shrink() decides to not shrink grants - a client is
supposed to restore its cl_grant_avail in osc_update_grant(). In case
of read OBD_MD_FLGRANT is not set on reply's body->oa.o_valid, so
osc_update_grant() misses the cl_grant_avail update. As result server
keeps thinking that client has a lot of grants while a client thinks
that it is missing grants badly. That may lead to performance
degradation.
A test to illustrate the issue is included.
Test-Parameters: trivial testlist=sanity
Change-Id: Ibe7ce0af5701226c8be3ae3f9ad57c354791fa0f
Signed-off-by: Vladimir Saveliev <vlaidimir.saveliev@hpe.com>
HPE-bug-id: LUS-9943
Reviewed-on: https://review.whamcloud.com/43375
Tested-by: jenkins <devops@whamcloud.com>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-by: Mike Pershin <mpershin@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Alexander Zarochentsev [Thu, 12 Nov 2020 16:47:36 +0000 (19:47 +0300)]
LU-15146 mdt: mdt_lvb2reply crash fix
mdt_lvb2body may crash if res->lr_lvb_data is
not allocated, make it tolerant to not initialized
lvb_data pointer.
HPE-bug-id: LUS-9549
Signed-off-by: Alexander Zarochentsev <alexander.zarochentsev@hpe.com>
Change-Id: Ie31cbba9187f9b04b3d1f8d2abc59e0612a44b41
Reviewed-on: https://review.whamcloud.com/45334
Reviewed-by: Mike Pershin <mpershin@whamcloud.com>
Tested-by: jenkins <devops@whamcloud.com>
Reviewed-by: Andrew Perepechko <andrew.perepechko@hpe.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Sebastien Buisson [Mon, 18 Oct 2021 11:26:40 +0000 (13:26 +0200)]
LU-15047 gss: gss integrity check with multi-rail
With multi-rail, a primary NID is used as node identifier, but LNet
decides which NID is actually used for sending/receiving data, on a
per request basis.
For the integrity check mechanism implemented as part of GSS, the
primary NID must be used in order to compute HMAC with the correct
key, independently of the actual NID for the current request.
Signed-off-by: Sebastien Buisson <sbuisson@ddn.com>
Change-Id: I2bf3974d3aa0e8365a9413dca56c69ee3734c12b
Reviewed-on: https://review.whamcloud.com/45277
Tested-by: jenkins <devops@whamcloud.com>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Jeremy Filizetti <jeremy.filizetti@gmail.com>
Reviewed-by: James Simmons <jsimmons@infradead.org>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Lei Feng [Tue, 12 Oct 2021 06:33:22 +0000 (14:33 +0800)]
LU-14587 ptlrpc: remove LASSERT in nrs_polices proc handler
It's not necessary to LASSERT() in nrs_polices proc handler.
CERROR() and returning error is good enough.
Signed-off-by: Lei Feng <flei@whamcloud.com>
Test-Parameters: trivial
Change-Id: I09f06dc4ab90e49b2df66a9b47a74678c64cdd2f
Reviewed-on: https://review.whamcloud.com/45200
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Li Xi <lixi@ddn.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Serguei Smirnov [Thu, 21 Oct 2021 21:03:34 +0000 (14:03 -0700)]
LU-15089 tests: allow enough time to create tcp connections
Allow enough time to create tcp connections before counting them
when testing socklnd conns_per_peer setting in sanity-lnet test_230
Test-Parameters: trivial testlist=sanity-lnet
Fixes:
a5cbe7883db6 ("LU-12815 socklnd: allow dynamic setting of conns_per_peer")
Signed-off-by: Serguei Smirnov <ssmirnov@whamcloud.com>
Change-Id: Ia3e25de157da03d6129b108c1af9632a8faf8efd
Reviewed-on: https://review.whamcloud.com/45331
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Chris Horn <chris.horn@hpe.com>
Tested-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Chris Horn [Wed, 20 Oct 2021 19:58:52 +0000 (14:58 -0500)]
LU-15125 o2iblnd: wrong list used for kib_connd_waits
The ibc_list field of struct kib_conn is used for the kib_connd_waits
list, but kiblnd_connd() uses ibc_sched_list in the
list_first_entry_or_null macro.
Test-Parameters: trivial
Fixes:
34b57a6f8f ("LU-12678 lnet: use list_first_entry() in lnet/klnds subdirectory.")
Signed-off-by: Chris Horn <chris.horn@hpe.com>
Change-Id: I0cc8a94636a5129956c53e48ae96b27ece5f0228
Reviewed-on: https://review.whamcloud.com/45316
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: James Simmons <jsimmons@infradead.org>
Reviewed-by: Alexey Lyashkov <alexey.lyashkov@hpe.com>
Reviewed-by: Serguei Smirnov <ssmirnov@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Andriy Skulysh [Wed, 14 Oct 2020 09:01:51 +0000 (12:01 +0300)]
LU-15118 ldlm: no free thread to process resend request
MDS grants lock request and sends a reply,
input request queue can be filled immediately
with conflicting lock enqueue request.
So there isn't any free thread to process resend of
first lock enqueue request if the client has failed
to receive the reply.
Process lock enqueue resends with existing lock on MDS
in high priority queue.
Change-Id: If7b94690100b44c774dc14231ed4902f701ed807
HPE-bug-id: LUS-9444
Signed-off-by: Andriy Skulysh <c17819@cray.com>
Reviewed-by: Alexander Boyko <alexander.boyko@hpe.com>
Reviewed-by: Alexander Zarochentsev <c17826@cray.com>
Reviewed-by: Vitaly Fertman <c17818@cray.com>
Reviewed-by: Andrew Perepechko <c17827@cray.com>
Reviewed-on: https://review.whamcloud.com/45272
Tested-by: jenkins <devops@whamcloud.com>
Reviewed-by: Andrew Perepechko <andrew.perepechko@hpe.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Andreas Dilger [Thu, 14 Oct 2021 20:01:30 +0000 (14:01 -0600)]
LU-15106 ofd: quiet deprecated param warning
There are a number of obdfilter parameter files that report a
warning even when they are read, which is confusing for users
if there is a tool that is scraping all available parameters:
# lctl get_param obdfilter.*.*
ofd: 'obdfilter.*.read_cache_enabled' is deprecated,
use 'osd-*.read_cache_enabled' instead
ofd: 'obdfilter.*.readcache_max_filesize' is deprecated,
use 'osd-*.readcache_max_filesize' instead
ofd: 'obdfilter.*.sync_on_lock_cancel' is deprecated,
use 'obdfilter.*.sync_lock_cancel' instead
ofd: 'obdfilter.*.writethrough_cache_enabled' is deprecated,
use 'osd-*.writethrough_cache_enabled' instead
It should only print a message if the parameters are actually written.
Also fix the messages to reference the correct parameter names.
Most of these parameter links were added in 2.4 with the addition of
osd-ldiskfs. However, the deprecation warnings were only added in
2.12.53 and slated for removal in 2.15, but were not backported to
2.12 LTS, and there hasn't been an LTS release since then, so it is
better bump removal so the upcoming 2.15 LTS release includes them.
Fix the test scripts to only use the new parameter names, to avoid
spurious warning messages. We don't test interop with 2.3 anymore.
Test-Parameters: trivial
Fixes:
7df7347b7b18 ("LU-12967 ofd: restore sync_on_lock_cancel tunable")
Fixes:
493cd8088388 ("LU-8066 osd: migrate from proc to sysfs")
Signed-off-by: Andreas Dilger <adilger@whamcloud.com>
Change-Id: Ie548e5b6af5463959fb4774e31996097373ebbe5
Reviewed-on: https://review.whamcloud.com/45246
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Alex Zhuravlev <bzzz@whamcloud.com>
Reviewed-by: Patrick Farrell <pfarrell@whamcloud.com>
Reviewed-by: James Simmons <jsimmons@infradead.org>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Elena Gryaznova [Thu, 14 Oct 2021 11:03:37 +0000 (14:03 +0300)]
LU-15104 tests: create client dirs with custom stripe count
Random ha_dir_stripe_count allows clients to create directories
with different stripe counts in one ha.sh test session.
Please use
DSTRIPECOUNT=3 DSTRIPECOUNTRAND=true
to create the directories with stripe counts 1,2 or 3.
Test-Parameters: trivial
Signed-off-by: Elena Gryaznova <elena.gryaznova@hpe.com>
HPE-bug-id: LUS-10435
Reviewed-by: Vladimir Saveliev <vlaidimir.saveliev@hpe.com>
Reviewed-by: Andriy Skulysh <andriy.skulysh@hpe.com>
Change-Id: I0092958a9bcf1991adc9f45ac1fbed9340a06c57
Reviewed-on: https://review.whamcloud.com/45240
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Jian Yu [Wed, 13 Oct 2021 07:12:03 +0000 (00:12 -0700)]
LU-14195 obdclass: change list_sort to use const pointers
Kernel 5.10.70 commit
4f0f586bf0c898233d8f316f471a21db2abd522d
defines the list_cmp_func_t type and changes the comparison
function types of all list_sort() callers to use const pointers
to avoid type mismatches.
Change-Id: I40d37ec0f0d485d0deebaa9dc3493f2865f76ec9
Signed-off-by: Jian Yu <yujian@whamcloud.com>
Reviewed-on: https://review.whamcloud.com/45219
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Tested-by: jenkins <devops@whamcloud.com>
Reviewed-by: James Simmons <jsimmons@infradead.org>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Andrew Perepechko [Mon, 11 Oct 2021 19:11:05 +0000 (22:11 +0300)]
LU-15081 vfs: set_nlink() is not race-safe
set_nlink() is not atomic wrt race with itself and
the following warning may be triggered by VFS:
WARNING: CPU: 5 PID: 195090 at fs/inode.c:241 __destroy_inode+0xdb/0xf0
It does not seem important what exact nlink value is the result
of the race. However, we need to protect the superblock remove
counter.
Signed-off-by: Andrew Perepechko <andrew.perepechko@hpe.com>
HPE-bug-id: LUS-9825
Change-Id: I67bc345b9a9e43fb88d919a83246759d11604b03
Reviewed-on: https://review.whamcloud.com/45191
Tested-by: jenkins <devops@whamcloud.com>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-by: Alexander Boyko <alexander.boyko@hpe.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Andreas Dilger [Thu, 30 Sep 2021 02:51:51 +0000 (20:51 -0600)]
LU-15045 utils: fix lfs_migrate for files with spaces
Fix the lfs_migrate script to properly quote "$OLDNAME" so that
it works for filenames with spaces and other characters in them.
Test-Parameters: trivial
Fixes:
8bedfa377fbd ("LU-11510 lfs: migrate a composite layout file correctly")
Fixes:
128137adfc53 ("LU-13090 utils: fix lfs_migrate -p for file with pool")
Signed-off-by: Andreas Dilger <adilger@whamcloud.com>
Change-Id: Ic00f41f3a91ad9dfa491ff57768a3da0c6300c1e
Reviewed-on: https://review.whamcloud.com/45173
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Arshad Hussain <arshad.hussain@aeoncomputing.com>
Reviewed-by: Olaf Faaland-LLNL <faaland1@llnl.gov>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Alexander Boyko [Thu, 7 Oct 2021 09:57:41 +0000 (05:57 -0400)]
LU-15071 utils: tunefs erease-params for zfs
The patch exclude special zfs params for tunefs erase-params,
skip nvlist modifying. Also fixes test_89 conf-sanity.
tunefs --erase-params produced segmentation fault with old code.
Test-Parameters: trivial fstype=zfs testlist=conf-sanity
HPE-bug-id: LUS-10314
Signed-off-by: Alexander Boyko <alexander.boyko@hpe.com>
Change-Id: Ic8385a99ca896ce6d855692b3f77e198bf583d94
Reviewed-on: https://review.whamcloud.com/45145
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Andriy Skulysh <andriy.skulysh@hpe.com>
Reviewed-by: Artem Blagodarenko <artem.blagodarenko@hpe.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Sergey Cheremencev [Wed, 23 Jun 2021 15:08:00 +0000 (18:08 +0300)]
LU-15067 lod: fix error handling in lod_new_pool
- correct error handling in lod_new_pool - ENOMEM
from tgt_pool_init may cause incorrect pool_count.
- optimisation in lu_tgt_pool_add. Do not extend
a pool, if the target is already exists.
HPE-bug-id: LUS-6995
Change-Id: I0ea00d335217f3994334bd02ae36c34653e7da98
Reviewed-by: Alexander Zarochentsev <c17826@cray.com>
Reviewed-by: Artem Blagodarenko <artem.blagodarenko@hpe.com>
Signed-off-by: Sergey Cheremencev <sergey.cheremencev@hpe.com>
Reviewed-on: https://review.whamcloud.com/45137
Tested-by: jenkins <devops@whamcloud.com>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Andreas Dilger [Wed, 15 Sep 2021 23:30:23 +0000 (17:30 -0600)]
LU-15011 lod: count all spilling events
when target pool is used to as the original has no enough space.
lctl lod.*.pool.<poolname>.spill_hit can be used to get the counter.
Signed-off-by: Alex Zhuravlev <bzzz@whamcloud.com>
Change-Id: I6d54a2b910705da182b5f4118e535d196cdab004
Reviewed-on: https://review.whamcloud.com/44947
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: John L. Hammond <jhammond@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Lai Siyao [Tue, 7 Sep 2021 09:33:21 +0000 (05:33 -0400)]
LU-13076 dne: dir migrate in QOS mode
Support "lfs migrate -m -1 ..." to migrate directory to MDTs by
space and inode usage, if system is balanced, the target MDT is
chosen in roundrobin mode, otherwise the less full MDTs will be
chosen, and the most full MDT is avoided.
Another minor change: if directory is migrated to specific MDTs,
and the target stripe count is more than 1, its subdirs may not be
migrated to the specified MDT in the command, but migrated to the
MDT where its parent stripe is located (subdir will be striped too),
as can avoid unnecessary remote directories. NB, for command like
"lfs migrate -m 0,1,2 ...", though the subdir may be located on
either MDT0, MDT1 or MDT2, its stripes will be striped over these
three MDTs, but for command like "lfs migrate -m 0 -c 3...", the
subdir may be striped on other MDTs if the subdir is not located on
MDT0.
Add sanity 230u.
Signed-off-by: Lai Siyao <lai.siyao@whamcloud.com>
Change-Id: I6e9c3d75bfc240b21c65ba27cd5e4bcca7058325
Reviewed-on: https://review.whamcloud.com/44886
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-by: Hongchao Zhang <hongchao@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
James Simmons [Mon, 11 Oct 2021 15:10:04 +0000 (11:10 -0400)]
LU-14927 quota: move qsd_transfer to lquota module
With osd-zfs inheriting the tainted state of ZFS we can no longer
directly use kernel internals that exported GPL only. The osd-zfs
modules quota code uses some of these internal functions. The
osd-zfs qsd_transfer() function is generic enough that we can
move it to the lquota modules.
Change-Id: I735db0266306477cd5558968f1f3ed1f6e1b32da
Signed-off-by: James Simmons <jsimmons@infradead.org>
Reviewed-on: https://review.whamcloud.com/44735
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Tested-by: jenkins <devops@whamcloud.com>
Reviewed-by: Alex Zhuravlev <bzzz@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Alex Zhuravlev [Mon, 16 Aug 2021 17:22:00 +0000 (20:22 +0300)]
LU-14938 tests: fail_abort() in t-f to take care of MDTs
fail_abort() in test-framework ensures that the clients
are back after evictions. the same should be done for
MDTs as otherwise any subsequent test may fail due to
another MDT observing eviction and interrupting current
request with -EIO.
Signed-off-by: Alex Zhuravlev <bzzz@whamcloud.com>
Change-Id: I0a00ece52d28c6d28eef029a4f87a348efaa041c
Reviewed-on: https://review.whamcloud.com/44671
Reviewed-by: Mike Pershin <mpershin@whamcloud.com>
Tested-by: jenkins <devops@whamcloud.com>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Sebastien Buisson [Fri, 13 Aug 2021 15:16:08 +0000 (17:16 +0200)]
LU-14937 build: re-use config cache in 'make rpms/debs'
Idea is to pass along the value of the -C or --cache-file options from
the initial ./configure to the one launched as part of the rpm or deb
build.
But all the environment variables related cache info must be removed
otherwise the config cache file cannot be reused.
Test-Parameters: trivial
Signed-off-by: Sebastien Buisson <sbuisson@ddn.com>
Change-Id: Iab4ae2815961ba10132d9cb44f82ca58d313e908
Reviewed-on: https://review.whamcloud.com/44659
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Shaun Tancheff <shaun.tancheff@hpe.com>
Reviewed-by: Patrick Farrell <pfarrell@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Lai Siyao [Tue, 3 Aug 2021 15:56:47 +0000 (11:56 -0400)]
LU-14909 test: mkdir_on_mdt0 to mkdir on MDT0
Many sub tests in recovery-small.sh and replay-single.sh need to mkdir
on MDT0, use mkdir_on_mdt0() to create such directories.
Test-Parameters: trivial mdscount=2 mdtcount=4 testlist=recovery-small.sh,replay-single.sh
Signed-off-by: Lai Siyao <lai.siyao@whamcloud.com>
Change-Id: Id4a44da062350ea284f51c8c821302aebbfe9dee
Reviewed-on: https://review.whamcloud.com/44544
Reviewed-by: Alex Zhuravlev <bzzz@whamcloud.com>
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Hongchao Zhang [Wed, 30 Jun 2021 11:15:03 +0000 (19:15 +0800)]
LU-14831 osd-ldiskfs: uninited osd_inode_id is used
In osd_fid_lookup, the "osd_inode_id" could be used uninitializedly
if the FID doesn't exist in OI, which cause some faked FID/inode
pair to be inserted into OI file and the OI scrub thread could be
triggered repeatedly.
Change-Id: I9100dece9e94d3e590f17fb4498601876aa1edaa
Signed-off-by: Hongchao Zhang <hongchao@whamcloud.com>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-on: https://review.whamcloud.com/44349
Tested-by: jenkins <devops@whamcloud.com>
Reviewed-by: Li Dongyang <dongyangli@ddn.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Mikhail Pershin [Fri, 4 Dec 2020 12:06:37 +0000 (15:06 +0300)]
LU-14174 lfs: llapi_mirror_find() signed return
Check return codes from llapi_mirror_find() as signed value
to don't count errors as valid mirror ids
Signed-off-by: Mikhail Pershin <mpershin@whamcloud.com>
Change-Id: I9754eaa063c9a2d07d8b815a86e7597922201f9c
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-on: https://review.whamcloud.com/41757
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: John L. Hammond <jhammond@whamcloud.com>
Shaun Tancheff [Tue, 24 Aug 2021 08:25:32 +0000 (03:25 -0500)]
LU-14409 ldiskfs: Add support for SUSE 5.3.18-24.46.1
Linux-commit:
f902b216501094495ff75834035656e8119c537f
ext4: fix bogus warning in ext4_update_dx_flag()
The update breaks the ldiskfs pdirop patch which disables
ext4_update_dx_flag.
SUSE 5.3.18-24.46 can directly use the 5.4.0-66-ubuntu20.series
Test-Parameters: trivial
HPE-bug-id: LUS-9684, LUS-9758
Signed-off-by: Shaun Tancheff <shaun.tancheff@hpe.com>
Change-Id: I9271ee80c05715d7dcec78535cfde1e384ba40e9
Reviewed-on: https://review.whamcloud.com/41473
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-by: James Simmons <jsimmons@infradead.org>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Elena Gryaznova [Thu, 14 Oct 2021 16:18:31 +0000 (19:18 +0300)]
LU-12391 tests: mdsrate tests improvements
Patch improves mdsrate tests to work with
striped directories which are created if
MDSRATE_ENABLE_DNE=true.
mdsrate.c is fixed to not fail if --mdtcount
option is used and ndirs=1. Without this fix
mdsrate --mdcount
fails as:
cannot create stripe dir: File exists
when the ranks do system(mkdir_cmd) on the
directory which was already created by the
first executed rank.
Patch sets mdt.*.enable_remote_dir_gid to "-1"
to allow mpiuser to create remote directories.
Patch makes the mdsrate based tests a bit more
verbose: mdsrate create/mknod is called with
--debug option if VERBOSE set to "true".
Test-Parameters: trivial testlist=performance-sanity
Fixes:
f31c60c97328 ("LU-1187 tests: Add mntfmt/mntcount/mdtcount to mdsrate")
Signed-off-by: Elena Gryaznova <elena.gryaznova@hpe.com>
HPE-bug-id: LUS-7262
Reviewed-by: Vladimir Saveliev <vlaidimir.saveliev@hpe.com>
Reviewed-by: Alexander Boyko <alexander.boyko@hpe.com>
Change-Id: Ieb32ad7dfad838fc9124740236889a5fe47cb901
Reviewed-on: https://review.whamcloud.com/35069
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Andreas Dilger [Wed, 19 Sep 2018 21:08:47 +0000 (17:08 -0400)]
LU-11407 obdclass: add start time to stats files
When the stats files are initialized or reset, store the current
timestamp with the stats. That allows computing average IO and
RPC rates over the accumulated stats lifetime, in addition to the
normal incremental operation rates found by comparing successive
values read from the stats file with the read interval.
Any stats that currently print the "snapshot_time:" header will
now also print "start_time:" and "elapsed_time:" fields as well.
Consolodate this printing into a helper function instead of
duplicating very similar code in many different functions. Output
can't be exactly the same for all callers, because these fields are
embedded into different types of output files, but it is very close.
Change struct rename_stats and brw_stats to use a common name prefix.
Change the obd_job_stats timestamps to ktime_t so that we can use the
common helper function for printing the header. It is easier to store
ojs_cleanup_interval internally as 1/2 of the maximum stats age, since
since division is more easily done when the value is initially set as
seconds compared to when it is ktime_t. This may also be a tiny bit
more efficient since we don't do a divide/shift on each access.
Signed-off-by: Andreas Dilger <adilger@whamcloud.com>
Change-Id: Iacefa17def455ef53a28fd14b6d4c670463ebbe5
Reviewed-on: https://review.whamcloud.com/33201
Tested-by: jenkins <devops@whamcloud.com>
Reviewed-by: Ben Evans <beevans@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Arshad Hussain <arshad.hussain@aeoncomputing.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
git://git.whamcloud.com / fs / lustre-release.git / log