Whamcloud - gitweb
Chris Horn [Fri, 10 Jul 2020 17:33:50 +0000 (12:33 -0500)]
LU-13779 lnet: Correct asymmetric route detection
Failure to lookup the remote net for LNET_NIDNET(src_nid) indicates an
asymmetric route, but we do not drop the message in this case. Another
problem with this code is that there is no guarantee that we'll have a
route->lr_lnet that matches the net of ni->ni_nid.
We can move the asymmetric route detection to after we have looked up
the lpni of from_nid. Then, we can look at just the routes associated
with the gateway that owns the lpni. If one of those routes has
lr_net == LNET_NIDNET(src_nid), then the route is symmetrical.
Fixes:
4932febc12 ("LU-11894 lnet: check for asymmetrical route messages")
Test-Parameters: trivial
HPE-bug-id: LUS-9087
Signed-off-by: Chris Horn <chris.horn@hpe.com>
Change-Id: I8044d3f53e6f000c1e4d7c4e34b3b21afe0f9711
Reviewed-on: https://review.whamcloud.com/39349
Reviewed-by: Neil Brown <neilb@suse.de>
Reviewed-by: Sebastien Buisson <sbuisson@ddn.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>
Mr NeilBrown [Mon, 6 Jul 2020 12:34:41 +0000 (08:34 -0400)]
LU-12678 socklnd: change various ints to bool.
Each of these int variables, and one int function, are
really truth values, so change to bool.
Also convert some spaces to tabs etc.
Signed-off-by: Mr NeilBrown <neilb@suse.de>
Change-Id: Ia62a86e549c90a287a20a3b2ef7533c1b700d17e
Reviewed-on: https://review.whamcloud.com/39302
Tested-by: jenkins <devops@whamcloud.com>
Reviewed-by: James Simmons <jsimmons@infradead.org>
Reviewed-by: Chris Horn <chris.horn@hpe.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Mr NeilBrown [Fri, 3 Jul 2020 03:51:50 +0000 (13:51 +1000)]
LU-13783 libcfs: support removal of kernel_setsockopt()
Linux 5.8 removes kernel_setsockopt() and kernel_getsockopt(), and
provides some helper functions for some accesses that are
not trivial.
This patch adds those helpers to libcfs when they are not available,
and changes (nearly) all calls to kernel_[gs]etsockopt() to
either use direct access to a helper call.
->keepalive() is not available before v4.11-rc1~94^2~43^2~14
and there is no helper function, so for SO_KEEPALIVE we
need to have #ifdef code in the C file.
TCP_BACKOFF* setting are not converted as they are not available in
any upstream kernel, so no conversion is possible.
Also include some minor style fixes and change lnet_sock_setbuf() and
lnet_sock_getbuf() to be 'void' functions.
Test-Parameters: trivial
Signed-off-by: Mr NeilBrown <neilb@suse.de>
Change-Id: I539cf8d20555ddb3565fa75130fdd3acf709c545
Reviewed-on: https://review.whamcloud.com/39259
Tested-by: jenkins <devops@whamcloud.com>
Reviewed-by: Aurelien Degremont <degremoa@amazon.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Chris Horn <chris.horn@hpe.com>
Reviewed-by: James Simmons <jsimmons@infradead.org>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Alex Zhuravlev [Fri, 5 Jun 2020 12:15:22 +0000 (15:15 +0300)]
LU-13636 obdclass: drop nlink if directory is removed
To make e2fsck happy. Otherwise, all the features using
local directories (quota, nodemap, nid tables) can leave
orphaned objects as nlink doesn't drop to 0.
Signed-off-by: Alex Zhuravlev <bzzz@whamcloud.com>
Change-Id: I9e20a304d66c61f312168715e888757bc06b6ed0
Reviewed-on: https://review.whamcloud.com/38844
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>
Mr NeilBrown [Mon, 28 Dec 2020 20:56:12 +0000 (15:56 -0500)]
LU-12514 llite: move client mounting from obdclass to llite
Mounting a lustre client is currently handled
in obdclass, using services from llite.
This requires obdclass to load the llite module
and set up inter-module linkage.
The purpose of this was for common code to support both
client and server mounts. This isn't really a good idea
and need to go. For lustre servers we already use a
separate filesystem type.
So move the mounting code from obdclass/obd_mount to llite/super25
and remove the inter-module linkages.
Add some EXPORT_SYMBOL() so that llite can access some helpers
that remain in obdclass.
Linux-commit:
a989830c88149511ee840356d9c1b34304bac576
Change-Id: Ia33bd55a042f90b178156c745a8072b516f00568
Signed-off-by: Mr NeilBrown <neilb@suse.de>
Reviewed-on: https://review.whamcloud.com/37693
Tested-by: Maloo <maloo@whamcloud.com>
Tested-by: jenkins <devops@whamcloud.com>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Mr NeilBrown [Mon, 24 Aug 2020 22:28:18 +0000 (08:28 +1000)]
LU-11085 mdt: revise recording of hsm progress updates.
When copy tool is migrating a file for HSM purposes it can report
progress as individual intervals, and the total covered by all the
intervals can be requested.
This patch makes various changes to the code for recording the
intervals.
- switch to the Linux interval-tree implementation rather than the
lustre one.
- detect overlapping intervals as well as duplicates. Any overlapping
or adjacent intervals are removed and the space which they covered
is added the the new interval.
- keep track of the total of all current intervals, so that it can be
returned on request without needing to examine the interval tree.
- use a spinlock rather than a mutex to protect against parallel
updates, as all operations are non-blocking.
Also add a test to llapi_hsm_test.c to send overlapping intervals and
check the result is as expected.
Signed-off-by: Mr NeilBrown <neilb@suse.de>
Change-Id: Id300b64949a3416ee3282c5e4ce82122c9e4e2f0
Reviewed-on: https://review.whamcloud.com/39725
Reviewed-by: Quentin Bouget <quentin.bouget@cea.fr>
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>
Mr NeilBrown [Thu, 17 Dec 2020 23:18:48 +0000 (10:18 +1100)]
LU-6142 lod: remove unnecessary variables
Both lprocfs_register() and dt_tunables_init() treat a NULL pointer as
being equivalent to an empty array.
So discard the empty arrays and use the NULL pointer.
Signed-off-by: Mr NeilBrown <neilb@suse.de>
Change-Id: I620fbf1c771ee80d2ddff0f38a87c2f08bae0e4d
Reviewed-on: https://review.whamcloud.com/41766
Reviewed-by: James Simmons <jsimmons@infradead.org>
Reviewed-by: Arshad Hussain <arshad.hussain@aeoncomputing.com>
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Mr NeilBrown [Mon, 2 Nov 2020 00:44:02 +0000 (11:44 +1100)]
LU-6142 llite: move acl code into separate file.
acl code is subject to conditional compilation (only if fs acls are
enabled), so move it to a separate file.
Signed-off-by: Mr NeilBrown <neilb@suse.de>
Change-Id: I472fa80193ee47abbab857bcc6dd021ed42ae9a5
Reviewed-on: https://review.whamcloud.com/40829
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: Oleg Drokin <green@whamcloud.com>
Alexander Boyko [Tue, 24 Nov 2020 09:05:36 +0000 (04:05 -0500)]
LU-13974 tests: update log corruption
Test case reproduce missing object for sub transaction during
set xattr operation.
First setattr got -2, second already started, but didn't
make llog_add yet. In this case llog osp object is stale after
top_trans_start. So declaration phase can not refresh llogs. And
at llog_osd_write_rec osp object changes stale state to
valid(dt_attr_get), but llog handle and llog header are invalid.
A new record would be added to updatelog with wrong index.
In that case processing of update log fails with
fs1-MDT0001-osp-MDT0003: [0x2:0x400024d0:0x2] Invalid record: index
112926 but expected 112925
lod_sub_recovery_thread()) fs1-MDT0001-osp-MDT0003 get update log
failed: rc = -34
Recovery aborted, and clients are evicted.
HPE-bug-id: LUS-9030
Test-Parameters: testlist=sanity envdefinitions=ONLY="427"
Signed-off-by: Alexander Boyko <alexander.boyko@hpe.com>
Change-Id: I6a47fed1bc01f4be62216d1d0787adc413df0cf5
Reviewed-on: https://review.whamcloud.com/40743
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Alexander Zarochentsev <alexander.zarochentsev@hpe.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Chris Horn [Thu, 6 Aug 2020 21:24:57 +0000 (16:24 -0500)]
LU-13883 lnet: Lookup lpni after discovery
The lpni for a nid can change as part of the discovery process (see
lnet_peer_add_nid()). As such, callers of lnet_discover_peer_locked()
need to lookup the lpni again after discovery completes to make sure
they get the correct peer.
An exception is lnet_check_routers() which doesn't do anything with
the peer or peer NI after the call to lnet_discover_peer_locked().
If the router list is changed then lnet_check_routers() will already
repeat discovery.
HPE-bug-id: LUS-9167
Signed-off-by: Chris Horn <chris.horn@hpe.com>
Change-Id: I8bdfcb957e87f65ce65bfad81858a4ce3362298e
Reviewed-on: https://review.whamcloud.com/39747
Reviewed-by: Serguei Smirnov <ssmirnov@whamcloud.com>
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: James Simmons <jsimmons@infradead.org>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Chris Horn [Thu, 6 Aug 2020 21:39:27 +0000 (16:39 -0500)]
LU-13894 lnet: Transfer disc src NID when merging peers
If we're merging two peers in lnet_peer_data_present() then we need
to transfer the src NID stored in the peer whose ping buffer we are
processing to the peer that actually owns the NIDs in the ping
buffer. Otherwise it is possible that the subsequent push to the peer
that is being discovered will go out over an interface that the peer
does not know about and it will be dropped.
Test-Parameters: trivial
HPE-bug-id: LUS-9193
Signed-off-by: Chris Horn <chris.horn@hpe.com>
Change-Id: I050c7c1c2c0eddb8d5ff12f40342a8a02efacb9c
Reviewed-on: https://review.whamcloud.com/39607
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Serguei Smirnov <ssmirnov@whamcloud.com>
Reviewed-by: James Simmons <jsimmons@infradead.org>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Chris Horn [Thu, 6 Aug 2020 21:21:29 +0000 (16:21 -0500)]
LU-13895 lnet: Prevent discovery on deleted peer
We needn't perform any discovery activities on a peer that has had
lnet_peer_del() called on it.
Test-Parameters: trivial
HPE-bug-id: LUS-9192
Signed-off-by: Chris Horn <chris.horn@hpe.com>
Change-Id: I5c89dc89038d2c8bf4d2a29029af7720963b81a2
Reviewed-on: https://review.whamcloud.com/39605
Reviewed-by: Serguei Smirnov <ssmirnov@whamcloud.com>
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: James Simmons <jsimmons@infradead.org>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Chris Horn [Fri, 7 Aug 2020 16:02:10 +0000 (11:02 -0500)]
LU-13895 lnet: Prevent discovery on peer marked deletion
If a peer has been marked for deletion then we needn't perform any
other discovery operation on it. Integrate this peer state into the
top level of the discovery state machine so that it is checked before
any other state.
HPE-bug-id: LUS-9192
Signed-off-by: Chris Horn <chris.horn@hpe.com>
Change-Id: Ie9de5b0d38d720f4f49d7e4a0673a6b52f9d3d80
Reviewed-on: https://review.whamcloud.com/39604
Reviewed-by: Serguei Smirnov <ssmirnov@whamcloud.com>
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: James Simmons <jsimmons@infradead.org>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Chris Horn [Tue, 23 Jun 2020 18:02:51 +0000 (13:02 -0500)]
LU-13708 lnet: lnet_notify sets route aliveness incorrectly
lnet_notify() modifies route aliveness in two ways:
1. By setting lp_alive field of the lnet_peer struct.
2. By setting lr_alive field of the lnet_route struct (via call to
lnet_set_route_aliveness())
In both cases, the aliveness value assigned is determined by a call
to lnet_is_peer_ni_alive(), but that value only reflects the aliveness
of a particular peer NI. A gateway may have multiple peer NIs, so the
aliveness of a gateway peer (lp_alive) is not necessarily equivalent
to the aliveness of one of its NIs. Furthermore, the lr_alive field
is only used to determine route aliveness for path selection if
discovery is disabled locally or on the gateway (see
lnet_find_route_locked() and lnet_is_route_alive()).
In general, we should not set lp_alive based on an lnet_notify()
call, and we should only set lr_alive if discovery is disabled. For
lr_alive specifically, we should only set it for those routes that
have the peer NI as a next-hop.
An exception to the above exists when the reset argument to
lnet_notify() is set. The gnilnd uses this flag in its calls to
lnet_notify() because gnilnd receives out-of-band notifications of
node up and down events. Thus, when gnilnd calls lnet_notify() we
actually know whether the gateway peer is up or down and we can set
lp_alive appropriately.
net lock/EX is held by other callers of lnet_set_route_aliveness, so
we do the same in lnet_notify().
Fixes:
e35be987da ("LU-12422 lnet: discovery off route state update")
Fixes:
ebc9835a97 ("LU-12941 lnet: Add peer level aliveness information")
Test-Parameters: trivial
HPE-bug-id: LUS-9034
Signed-off-by: Chris Horn <chris.horn@hpe.com>
Change-Id: I2927e5f5ef849e45c233c92d2a6deca765e496eb
Reviewed-on: https://review.whamcloud.com/39160
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>
Sergey Gorenko [Thu, 4 Mar 2021 12:33:16 +0000 (14:33 +0200)]
LU-14488 o2ib: Use rdma_connect_locked if it is defined
rdma_connect_locked() is added in the upstream kernel 5.10 and
MOFED-5.2-2. After that, it is not allowed to call rdma_connect()
in RDMA CM event handler; rdma_connect_locked() must be used
instead.
This commit adds configure checks to detect whether
rdma_connect_locked() is available and updates the event handler
to call the correct function.
Test-Parameters: trivial
Signed-off-by: Sergey Gorenko <sergeygo@nvidia.com>
Change-Id: I8068d04810bf6f0200292a55f3fdcea8c71d44c1
Reviewed-on: https://review.whamcloud.com/41887
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>
Lai Siyao [Tue, 9 Feb 2021 14:09:09 +0000 (22:09 +0800)]
LU-14405 mdt: read LMV with mdt_stripe_get()
mdt_path_current() reads LMV into mdt_thread_info.mti_xattr_buf,
whose size is static, and will return -ERANGE if LMV contains too
many stripes, instead it should call mdt_stripe_get(), the latter
will allocate dynamic memory for LMV.
Test-Parameters: mdtcount=8
Signed-off-by: Lai Siyao <lai.siyao@whamcloud.com>
Change-Id: I1ed78f7a7f951fa5984e604a8773143a70b419e7
Reviewed-on: https://review.whamcloud.com/41452
Tested-by: jenkins <devops@whamcloud.com>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Yingjin Qian <qian@ddn.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Andreas Dilger [Thu, 11 Feb 2021 21:22:58 +0000 (14:22 -0700)]
LU-14058 tests: handle more MDTs in sanity.sh
Fix up sanity.sh test_160 to handle configurations with more MDTs.
The "fnv_1a_64" hash is _relatively_ uniform and harder to break
under normal (ab)use, it doesn't leave totally entries balanced.
Even "all_chars" hash has a repeat MDT every handful of entries.
Since we need perfect balance across MDTs, use "lfs mkdir -i".
Fix a bug in test_160g that wasn't setting changelog_max_idle_indexes
properly for test systems with more than 4 MDTs.
Test-Parameters: trivial testlist=sanity env=ONLY=160,230 mdtcount=8
Fixes:
489afbe69d5b ("LU-13321 tests: force even DNE file distribution")
Signed-off-by: Andreas Dilger <adilger@whamcloud.com>
Change-Id: I08bf2274a00fe1c6e52ec1a55f50dc8662d354a9
Reviewed-on: https://review.whamcloud.com/41485
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: James Nunez <jnunez@whamcloud.com>
Reviewed-by: Lai Siyao <lai.siyao@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Sebastien Buisson [Wed, 9 Dec 2020 17:53:12 +0000 (18:53 +0100)]
LU-14204 tests: make sure we have a single import
In sanity, retrieve the exact name of the import being used on the
client, in order to properly get information such as lock_count or
lru_size.
Change-Id: I065b7da7990c7171d5baa24f3400c5f8ffc12fc9
Test-Parameters: trivial
Test-Parameters: env=SHARED_KEY=true testlist=sanity
Signed-off-by: Sebastien Buisson <sbuisson@ddn.com>
Reviewed-on: https://review.whamcloud.com/41758
Tested-by: jenkins <devops@whamcloud.com>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: James Nunez <jnunez@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
John L. Hammond [Tue, 23 Feb 2021 15:40:08 +0000 (09:40 -0600)]
LU-14468 utils: improve 'lfs rmfid' error messages
In lfs_rmfid_and_show_errors(), convert the error messages printed by
'lfs rmfid' from the format
rmfid([0x20001a9f5:0x159:0x0]): rc = -39
to
lfs rmfid: cannot remove [0x20001a9f5:0x155:0x0]: Directory not empty
Simplify the logic and swap rc and rc2 to follow conventions.
Signed-off-by: John L. Hammond <jhammond@whamcloud.com>
Change-Id: Iccd9e1054ed8842fc4f65dd601077cfdeaa1320c
Reviewed-on: https://review.whamcloud.com/41727
Tested-by: jenkins <devops@whamcloud.com>
Reviewed-by: Jian Yu <yujian@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Alex Zhuravlev <bzzz@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Jian Yu [Tue, 23 Feb 2021 17:26:37 +0000 (09:26 -0800)]
LU-14385 utils: add range check to strtoul() in lfs.c
Most of the strtoul() functions called in lfs.c
did not check the range of the return value.
This patch fixes those issues.
Change-Id: If1eb64750507b5fa4e22abe710e475e2f0032b4d
Signed-off-by: Jian Yu <yujian@whamcloud.com>
Reviewed-on: https://review.whamcloud.com/41726
Tested-by: jenkins <devops@whamcloud.com>
Reviewed-by: John L. Hammond <jhammond@whamcloud.com>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Sebastien Buisson [Mon, 22 Feb 2021 15:24:11 +0000 (00:24 +0900)]
LU-14462 gss: fix support for namespace in lgss_keyring
Fix the way lgss_keyring handles different mount namespaces,
so that we do not try to bind to a namespace that does not exist.
Fixes:
94c44c62de ("LU-7845 gss: support namespace in lgss_keyring")
Signed-off-by: Sebastien Buisson <sbuisson@ddn.com>
Change-Id: Ia5a5213399decc683d5e9401b6594e7fe579123f
Reviewed-on: https://review.whamcloud.com/41716
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: John L. Hammond <jhammond@whamcloud.com>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Lei Feng [Mon, 22 Feb 2021 02:06:03 +0000 (10:06 +0800)]
LU-13857 obdclass: Add white space to output valid YAML.
YAML needs a white space after the colon(:) between a pair of key and
value. In this case, if the integer is large enough, it will leave no
white space. So insert the white space forcefully.
Change-Id: I366b5399cc293a66a70ea6084c6a5fa30a58813b
Signed-off-by: Lei Feng <flei@whamcloud.com>
Reviewed-on: https://review.whamcloud.com/41709
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>
NeilBrown [Mon, 15 Feb 2021 14:57:07 +0000 (09:57 -0500)]
LU-9859 tracefile: convert list_for_each_entry_safe() to while(!list_empty())
These loops are removing all elements from a list.
So using while(!list_empty()) makes the intent clearer.
Linux-commit:
fdafb01e2c70e6b5321d158a2ff1f20a13d9b365
Change-Id: Idda25888e424a1deaa4d7c6fad427d494b1f56e5
Signed-off-by: NeilBrown <neilb@suse.de>
Reviewed-on: https://review.whamcloud.com/41670
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Yang Sheng <ys@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Jian Yu [Fri, 19 Feb 2021 23:33:34 +0000 (15:33 -0800)]
LU-14385 tests: verify lfs setstripe comp-flags and flags options
This patch adds more test cases to verify lfs setstripe
--comp-flags|--component-flags and --flags options.
Change-Id: Ie09089ceb65372fdf4e3b50df3771c9a355210cc
Signed-off-by: Jian Yu <yujian@whamcloud.com>
Reviewed-on: https://review.whamcloud.com/41423
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>
John L. Hammond [Wed, 3 Feb 2021 20:17:11 +0000 (14:17 -0600)]
LU-14398 lfs: use llapi_fid2path_at() in lfs_fid2path()
Use llapi_fid2path_at() in lfs_fid2path(). This avoids resolving and
opening the mount point for each FID argument passed. Make the -c,
--cur, --current option actually print the link. Add a more
descriptive long option name for this (--print-link). Update the
lfs-fid2path manpgae accordingly.
Signed-off-by: John L. Hammond <jhammond@whamcloud.com>
Change-Id: If851e4ce95f87d3188b644eb4a345ba3cfca530d
Reviewed-on: https://review.whamcloud.com/41407
Tested-by: jenkins <devops@whamcloud.com>
Reviewed-by: Jian Yu <yujian@whamcloud.com>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Vikentsi Lapa [Thu, 28 Jan 2021 16:08:38 +0000 (16:08 +0000)]
LU-14314 tests: skip test_16e from sanityn
To avoid test failure skip test_16e from sanityn
when Lustre version is below 2.13.53 on MDS server
Fixes:
92d799217aea ("LU-13227 sanityn 16a FAIL: fsx with O_DIRECT failed.")
Signed-off-by: Vikentsi Lapa <vlapa@whamcloud.com>
Change-Id: I562df7e02a9484fbc037597f012943cefa480fda
Test-Parameters: trivial testlist=sanityn
Reviewed-on: https://review.whamcloud.com/41353
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>
Alex Zhuravlev [Tue, 26 Jan 2021 14:18:59 +0000 (17:18 +0300)]
LU-14370 quota: use dt_sync() to flush pending writes
and find remaining quota.
Fixes:
18cd3e1e28 ("LU-12702 quota: wait pending write before acquiring remotely")
Signed-off-by: Alex Zhuravlev <bzzz@whamcloud.com>
Change-Id: I8f1ec334a1d1eefc385d8c6ef451de8a3f12365f
Reviewed-on: https://review.whamcloud.com/41322
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Tested-by: jenkins <devops@whamcloud.com>
Reviewed-by: Wang Shilong <wshilong@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Andreas Dilger [Sat, 9 Jan 2021 09:08:06 +0000 (02:08 -0700)]
LU-12125 mds: allow parallel regular file rename
Allow rename of non-directory files in the same directory to be done
in parallel, by only taking the DLM lock on the parent FID, without
also locking the global LUSTRE_BFL_FID (Big Filesystem Lock).
Older clients may not send the renamed file mode in mds_rec_rename.
In this case, the LUSTRE_BFL_FID lock will still be taken, and is not
worse than before parallel rename was allowed.
Similarly, if (for whatever reason) there is a mix of MDS versions
running in the same filesystem, at worst older MDSes will continue to
unnecessarily lock LUSTRE_BFL_FID before doing the file rename.
If MDT0000 is on an older MDS, but newer MDSes are doing renames of
non-directories, the newer MDSes will *not* lock LUSTRE_BFL_FID first,
but there will still be proper serialization from the parent directory
FID lock for other renames affecting the parent and the source/target
entries. That MDT0000 is unaware of the rename is the whole point.
In case of a race, where the file mode sent by the client is stale,
this is also not a concern, because the file mode is rechecked later
under lock and the rename fails if the source and target mode differ.
Test-Parameters: testlist=racer env=DURATION=3600
Signed-off-by: Andreas Dilger <adilger@whamcloud.com>
Change-Id: If330b53eb6db46e40f50fd7834a83e80db3ebbe5
Reviewed-on: https://review.whamcloud.com/41186
Tested-by: jenkins <devops@whamcloud.com>
Reviewed-by: Lai Siyao <lai.siyao@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Mike Pershin <mpershin@whamcloud.com>
Reviewed-by: Yingjin Qian <qian@ddn.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Andriy Skulysh [Wed, 19 Feb 2020 20:06:33 +0000 (22:06 +0200)]
LU-13991 ldlm: speedup flock reprocess
We can check for deadlock only for first
conflicting lock, the rest deadlock checks
will be performed after cancelation of
first conflicting lock.
Change-Id: I18359db405ab021a4f32ac833de203254097142d
HPE-bug-id: LUS-8509
Signed-off-by: Andriy Skulysh <c17819@cray.com>
Reviewed-by: Vitaly Fertman <c17818@cray.com>
Reviewed-by: Alexander Boyko <c17825@cray.com>
Tested-by: Alexander Lezhoev <c17454@cray.com>
Reviewed-on: https://review.whamcloud.com/40048
Tested-by: jenkins <devops@whamcloud.com>
Reviewed-by: Alexander Boyko <alexander.boyko@hpe.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Andreas Dilger [Fri, 18 Oct 2019 11:12:24 +0000 (20:12 +0900)]
LU-12885 mdd: clearly name variables for MAY_ flags
Clearly name variables for MAY_READ, MAY_WRITE, MAY_EXEC to
distinguish it from other "mask" variables.
The kernel VFS silently converts the MAY_READ, MAY_WRITE, and
MAY_EXEC flags to ACL_READ, ACL_WRITE, and ACL_EXECUTE modes for
the on-disk ACLs. It later also converts from the ACL_* flags
to POSIX rwx bits. Verify that these values are the same.
Test-Parameters: trivial
Signed-off-by: Andreas Dilger <adilger@whamcloud.com>
Change-Id: Idcd91d4c467c4415f1c67a5081721393cd3ebbe5
Reviewed-on: https://review.whamcloud.com/36520
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: James Simmons <jsimmons@infradead.org>
Reviewed-by: Arshad Hussain <arshad.hussain@aeoncomputing.com>
Reviewed-by: Neil Brown <neilb@suse.de>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Andreas Dilger [Fri, 18 Oct 2019 10:43:11 +0000 (19:43 +0900)]
LU-12885 llite: mark extended attr and inode flags
Clearly name the extended attribute and inode flags so that it is
possible to distinguish them more easily.
Test-Parameters: trivial
Signed-off-by: Andreas Dilger <adilger@whamcloud.com>
Change-Id: Id1ad2ba6411f69efb2de38e1019940f5fb3ebbe5
Reviewed-on: https://review.whamcloud.com/36519
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: James Simmons <jsimmons@infradead.org>
Reviewed-by: Arshad Hussain <arshad.hussain@aeoncomputing.com>
Reviewed-by: Neil Brown <neilb@suse.de>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Mr NeilBrown [Thu, 15 Oct 2020 22:43:09 +0000 (09:43 +1100)]
LU-6142 lustre: iput() can safely be passed NULL.
iput() is a no-op when passed a NULL pointer, so there is no
need to test for NULL before calling it - doing so clutters
the code.
Signed-off-by: Mr NeilBrown <neilb@suse.de>
Change-Id: Idcd1a6746ecc67dcfcb0713d2762ca0bdb29de19
Reviewed-on: https://review.whamcloud.com/40291
Tested-by: jenkins <devops@whamcloud.com>
Reviewed-by: Aurelien Degremont <degremoa@amazon.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: James Simmons <jsimmons@infradead.org>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Mr NeilBrown [Tue, 25 Aug 2020 00:03:17 +0000 (10:03 +1000)]
LU-11085 nodemap: switch interval tree to in-kernel impl.
Switch nodemap_range to use the in-kernel interval tree.
This has the same functionality, though often in a different form.
Signed-off-by: Mr NeilBrown <neilb@suse.de>
Change-Id: I7bf119bf8cd8f14dc66deb2736c2c97562bb0743
Reviewed-on: https://review.whamcloud.com/39724
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: James Simmons <jsimmons@infradead.org>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Shaun Tancheff [Sun, 30 Aug 2020 18:53:09 +0000 (13:53 -0500)]
LU-13485 libcfs: FIELD_SIZEOF macro removed
Linux v4.15-rc2-5-g4229a470175b introduced sizeof_field() macro
Linux v5.5-rc4-1-g1f07dcc459d5 removed FIELD_SIZEOF() macro
Provide a sizeof_field() macro in terms of FIELD_SIZEOF()
when sizeof_field() is not provided.
Test-Parameters: trivial
Signed-off-by: Shaun Tancheff <shaun.tancheff@hpe.com>
Change-Id: I48ca9abb931d58919d788199e5089984c9e854dd
Reviewed-on: https://review.whamcloud.com/39710
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-by: Sebastien Buisson <sbuisson@ddn.com>
Reviewed-by: James Simmons <jsimmons@infradead.org>
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Mr NeilBrown [Mon, 6 Jul 2020 12:34:40 +0000 (08:34 -0400)]
LU-12678 o2iblnd: convert peers hash table to hashtable.h
Using a hashtable.h hashtable, rather than bespoke code, has several
advantages:
- the table is comprised of hlist_head, rather than list_head, so
it consumes less memory (though we need to make it a little bigger
as it must be a power-of-2)
- there are existing macros for easily walking the whole table
- it uses a "real" hash function rather than "mod a prime number".
In some ways, rhashtable might be even better, but it can change the
ordering of objects in the table are arbitrary moments, and that could
hurt the user-space API. It also does not support the partitioned
walking that ksocknal_check_peer_timeouts() depends on.
Note that new peers are inserted at the top of a hash chain, rather
than appended at the end. I don't think that should be a problem.
Also various white-space cleanups etc.
Signed-off-by: Mr NeilBrown <neilb@suse.de>
Change-Id: I2917024835abdd327c7da11dee3fd369570a9671
Reviewed-on: https://review.whamcloud.com/39303
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Chris Horn <chris.horn@hpe.com>
Reviewed-by: James Simmons <jsimmons@infradead.org>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Mr NeilBrown [Mon, 6 Jul 2020 12:34:42 +0000 (08:34 -0400)]
LU-12678 lnet: discard LNET_MD_PHYS
This macro has no value and is never set.
It claims "compatibility with Cray Portals", yet cray-dvs
git://github.com/glennklockwood/cray-dvs.git
does not use it in any non-trivial way.
Much has changed in lnet and lib-md since 2007 when this
value was added - it seems likely that this really
is dead.
So remove it. If/when this results in problems, it can
easily be re-added and more details can be provided at
that time.
Test-Parameters: trivial testlist=sanity-lnet
Signed-off-by: Mr NeilBrown <neilb@suse.de>
Change-Id: Idef9389f4c0993adbdf088d0ccd9a0dc1449e86e
Reviewed-on: https://review.whamcloud.com/39301
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: Chris Horn <chris.horn@hpe.com>
Reviewed-by: Shaun Tancheff <shaun.tancheff@hpe.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Mr NeilBrown [Mon, 6 Jul 2020 12:34:43 +0000 (08:34 -0400)]
LU-12678 lnet: use init_wait() rather than init_waitqueue_entry()
init_waitqueue_entry(foo, current)
is equivalent to
init_wait(foo)
So use the shorter version.
Change-Id: Ic63e99d75986211d9655a89f56721394c7b3abb6
Signed-off-by: Mr. NeilBrown <neilb@suse.de>
Reviewed-on: https://review.whamcloud.com/39295
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Chris Horn <chris.horn@hpe.com>
Reviewed-by: James Simmons <jsimmons@infradead.org>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Mr NeilBrown [Mon, 6 Jul 2020 12:34:45 +0000 (08:34 -0400)]
LU-9859 libcfs: use wait_event_timeout() in tracefiled().
By using wait_event_timeout() we can make it more clear what is being
waited for, and when the loop terminates.
Signed-off-by: Mr NeilBrown <neilb@suse.de>
Change-Id: I5d015a3cbdbb342a5117e2c328680b3ec13aeb58
Reviewed-on: https://review.whamcloud.com/39293
Tested-by: jenkins <devops@whamcloud.com>
Reviewed-by: Shaun Tancheff <shaun.tancheff@hpe.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: James Simmons <jsimmons@infradead.org>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
James Simmons [Sat, 26 Dec 2020 17:19:19 +0000 (12:19 -0500)]
LU-12678 lnet: discard WIRE_ATTR
This macro adds nothing of value, and make the code harder to
read for new readers so it was remove for the Linux client.
We still want to keep track of what data structures are
transmitted over the wire and ensure the protocol does not get
broken. Move the wire protocol structures to their own header
files and add wire checking.
Linux-commit:
3e60455d1953fcee6042c107a8d5657886aa9c58
Test-Parameters: trivial
Change-Id: I4867e80bf8e8f0598d1920865d6f1b9ba920ce5b
Signed-off-by: Mr NeilBrown <neilb@suse.de>
Signed-off-by: James Simmons <jsimmons@infradead.org>
Reviewed-on: https://review.whamcloud.com/37914
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>
Shaun Tancheff [Thu, 10 Dec 2020 16:31:51 +0000 (10:31 -0600)]
LU-13239 ldiskfs: pass inode timestamps at initial creation
A previous patch https://github.com/Cray/lustre/commit/
6d4fb6694
"LUS-4880 osd-ldiskfs: pass uid/gid/xtime directly to ldiskfs"
was intended to be ported to upstream lustre but was lost.
The patch https://review.whamcloud.com/34685/
"LU-12151 osd-ldiskfs: pass owner down rather than transfer it"
passed the inode UID and GID down to ldiskfs at inode allocation
time to avoid the overhead of transferring quota from the inode
(initially created as root) over to the actual user of the file.
The two patches differed slightly in that the LUS-4880 included
passing the a/m/ctimes from osd-ldiskfs to ldiskfs at inode
creation time avoids overhead of setting the timestamps afterward.
Benchmarks using MDTEST:
mdtest -f 32 -l 32 -n 16384 -i 5 -p 120 -t -u -v -d mdtest
master patched
Operation Mean Std Dev Mean Std Dev
--------- ---- ------- ---- -------
Directory creation: 17008.593 72.700 17099.863 155.461
Directory stat : 170513.269 1456.002 170105.207 2349.934
Directory removal : 80796.147 2633.832 84480.222 892.536
File creation : 39227.419 7014.539 40429.900 6643.868
File stat : 101761.395 2979.802 103818.800 1146.689
File read : 86583.370 871.982 85725.254 965.862
File removal : 74923.504 761.048 75075.180 723.966
Tree creation : 588.570 244.534 608.332 123.939
Tree removal : 39.874 1.873 44.357 2.350
This patch also reorganizes the ldiskfs patch series in
order to accommodate struct iattr being added to
ldiskfs_create_inode.
All supported server platforms RHEL 7.5+, SUSE 12+ and
ubuntu 18+ are affected.
HPE-bug-id: LUS-7378, LUS-4880, LUS-8042, LUS-9157, LUS-8772, LUS-8769
Signed-off-by: Shaun Tancheff <shaun.tancheff@hpe.com>
Change-Id: I87e9c792b5240820bfd3a7268e477970ebac8465
Reviewed-on: https://review.whamcloud.com/37556
Reviewed-by: Petros Koutoupis <petros.koutoupis@hpe.com>
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Jian Yu <yujian@whamcloud.com>
Reviewed-by: Wang Shilong <wshilong@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Mr NeilBrown [Wed, 23 Oct 2019 00:30:50 +0000 (11:30 +1100)]
LU-12780 ofd: don't use ptlrpc_thread for consistency verification
The ofd module runs a consistency verification thread to verify parent
FID. Rather than using ptlrpc_thread to manage this, use native
kthreads functionality.
- startup-up code is moved out of the thread to before the
thread is started, which make error handling clearer.
As part of this, the lfsck_req_local struct is combined with
an lu_env and ofd_device pointer into a new oivm_args
which is passed to the thread a arguments - now it doesn't need
to allocate anything itself.
- Cleanup remains in the thread, so we add a completion to be
sure the thread has started before there is any chance of
kthread_stop() being called.
- kthread_stop() and kthread_should_stop() are used for stopping
the thread. wake_up_process() is used to wake it.
The thread sets TASK_IDLE at the top of the loop, and sets
TASK_RUNNING if anything is found to do. At the bottom of
the loop the 'schedule()' will only block if nothing was found
to be done.
Signed-off-by: Mr NeilBrown <neilb@suse.de>
Change-Id: Iec1de307ea48f7d26c60edf5d86eb0b7bf78f49a
Reviewed-on: https://review.whamcloud.com/36262
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>
Mr. NeilBrown [Thu, 19 Nov 2020 14:09:19 +0000 (09:09 -0500)]
LU-11085 ldlm: change lock_matches() to return bool.
The name of the function lock_matches() sounds like it
performs a test (it does) and so should return a bool.
Returning a bool gives a slight code simplification (in
search_queue) and more simplification in future patches.
Linux-commit:
e16983d96c775eb4527208d3c3d13f57e6d6233c
Change-Id: I1e3a09a0768abd0ab1cfada0fd69216cb9e85df7
Signed-off-by: Mr. NeilBrown <neilb@suse.de>
Reviewed-on: https://review.whamcloud.com/33854
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Mr NeilBrown [Tue, 24 Nov 2020 02:45:39 +0000 (13:45 +1100)]
LU-6142 libcfs: discard cfs_strrstr()
cfs_strrstr() is only used in one place, and it can easily be open
coded there without increasing code complexity. In particular the
fact that the "needle" cannot meaningfully be at the start of the
"haystack", means a simple loop does all we need.
In fact, there is room to improve the code in lwp_setup()
- sprintf isn't needed as the result is a constant that can
be calculated at compile time
- adding the nul termination is then not needed as the buffer
being copied to was initialised to zeroes.
Signed-off-by: Mr NeilBrown <neilb@suse.de>
Change-Id: I52b4abb36cf809d3bd9eebcc752959b0a81bfc13
Reviewed-on: https://review.whamcloud.com/40861
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Alex Zhuravlev <bzzz@whamcloud.com>
Reviewed-by: James Simmons <jsimmons@infradead.org>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Mr NeilBrown [Tue, 24 Nov 2020 02:28:01 +0000 (13:28 +1100)]
LU-6142 libcfs: discard cfs_firststr
The effect of cfs_firststr() can easily achieved with
skip_space() and strsep().
So use that instead.
Signed-off-by: Mr NeilBrown <neilb@suse.de>
Change-Id: Idcf8aa50b6aad052f7ee5341ce6d635495aa4990
Reviewed-on: https://review.whamcloud.com/40860
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>
Mr NeilBrown [Fri, 20 Nov 2020 02:20:23 +0000 (13:20 +1100)]
LU-6142 libcfs: discard PO2_ROUNDUP_TYPED, LOWEST_BIT_SET
LOWEST_BIT_SET() is never used.
PO2_ROUNDUP_TYPED() has the same function as 'round_up()'.
osd_roundup2blocksz() can be further simplified using
DIV_ROUND_UP_ULL().
Signed-off-by: Mr NeilBrown <neilb@suse.de>
Change-Id: I075328ec935eef49d3aeaf5ea1b79b943aadfa2e
Reviewed-on: https://review.whamcloud.com/40859
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Alex Zhuravlev <bzzz@whamcloud.com>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-by: James Simmons <jsimmons@infradead.org>
Mr NeilBrown [Thu, 15 Oct 2020 04:10:50 +0000 (15:10 +1100)]
LU-6142 llite: ll_lookup_finish_locks clean up
The 'inode' var can be used rather than dereferencing dentry->d_inode
a second time.
Also re-indent the function.
Signed-off-by: Mr NeilBrown <neilb@suse.de>
Change-Id: I16fb5480f7e2a8192a82e0c821597e14aad00b40
Reviewed-on: https://review.whamcloud.com/40830
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Lai Siyao <lai.siyao@whamcloud.com>
Reviewed-by: Aurelien Degremont <degremoa@amazon.com>
Reviewed-by: Arshad Hussain <arshad.hussain@aeoncomputing.com>
Reviewed-by: James Simmons <jsimmons@infradead.org>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Mr NeilBrown [Thu, 15 Oct 2020 03:32:02 +0000 (14:32 +1100)]
LU-6142 llite: don't cast arg to d_lustre_invalid()
There is no need to cast the arg to d_lustre_invalid() as it
is always of exactly the correct type.
Signed-off-by: Mr NeilBrown <neilb@suse.de>
Change-Id: I48b103abe2755a2d8892439974744dea232db515
Reviewed-on: https://review.whamcloud.com/40824
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>
Mr NeilBrown [Fri, 16 Oct 2020 03:29:32 +0000 (14:29 +1100)]
LU-6142 lustre: use is_root_inode()
Lustre has multiple tests to see if a given inode is the root of the
filesystem. Linux has (since 3.19) a helper function is_root_inode().
Use that throughout.
Signed-off-by: Mr NeilBrown <neilb@suse.de>
Change-Id: Ib5997e8976b6e935c268ad7ac9328fcd4d071b1f
Reviewed-on: https://review.whamcloud.com/40293
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-by: James Simmons <jsimmons@infradead.org>
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Mr NeilBrown [Thu, 15 Oct 2020 22:10:04 +0000 (09:10 +1100)]
LU-6142 llite: remove ll_dir_chain
ll_dir_chain is an empty structure that was introduced 13 years ago
and never used. As such it is clutter.
Remove it an all references to it. If the need which proposed it
ever does get addresses, having a clean slate in which to create
a suitable implementation is probably preferred.
Signed-off-by: Mr NeilBrown <neilb@suse.de>
Change-Id: I1eff80add8851ea11fff979f9b2d25916d6d3eed
Reviewed-on: https://review.whamcloud.com/40290
Tested-by: jenkins <devops@whamcloud.com>
Reviewed-by: Aurelien Degremont <degremoa@amazon.com>
Reviewed-by: James Simmons <jsimmons@infradead.org>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Mr NeilBrown [Thu, 15 Oct 2020 00:10:47 +0000 (11:10 +1100)]
LU-6142 lustre: remove non-static 'inline' markings.
There is rarely any point in marking a non-static function as
'inline'. The result is to compile a state-alone function that other
files can refer to, and also to inline the code where it is used in
the same file.
In many cases the non-static inline functions are not used in the same
file, so the 'inline' marking has no effect. In other cases it may
have an effect, but it can only be needed in highly performance
critical situations where a function call must be avoided, and that
doesn't seem like in any of these cases.
So just remove the "inline".
Test-Parameters: trivial
Signed-off-by: Mr NeilBrown <neilb@suse.de>
Change-Id: Ic3243ee80f9bfd75a67dd8c89ea07d08dc36425c
Reviewed-on: https://review.whamcloud.com/40289
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Aurelien Degremont <degremoa@amazon.com>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-by: James Simmons <jsimmons@infradead.org>
Mr NeilBrown [Thu, 27 Aug 2020 06:52:13 +0000 (16:52 +1000)]
LU-6142 lustre: convert snprintf to scnprintf as appropriate
The return value of snprintf() is the number of bytes that would have
been copies into the buffer if it was large enough.
Many places in the code use it as though it were the number of bytes
actually copied. In practice this (almost?) never makes a difference.
However it is poor style to use the wrong interfaces as it might one
day be copied to somewhere that it does make a difference.
So change these instances of snprintf to scnprintf which DOES return
the number of bytes actually copied.
This is all places where the return value is simply returned to the
call, and a couple of others.
Also change the declared buffer size in a couple of places to the
actual buffer size (PAGE_SIZE in these cases).
Test-Parameters: trivial
Signed-off-by: Mr NeilBrown <neilb@suse.de>
Change-Id: I5428760689eaca98e6bf2bceae119f1b7acf35ac
Reviewed-on: https://review.whamcloud.com/39744
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: James Simmons <jsimmons@infradead.org>
Reviewed-by: Shaun Tancheff <shaun.tancheff@hpe.com>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Mr NeilBrown [Thu, 27 Aug 2020 04:34:25 +0000 (14:34 +1000)]
LU-6142 lustre: mark strings in char arrays as const
Most array of strings are marked 'const', but the strings within the
arrays often aren't.
This patch mark all strings in const arrays of strings as const.
This allows them to be placed in read-only memory.
Test-Parameters: trivial
Signed-off-by: Mr NeilBrown <neilb@suse.de>
Change-Id: I93b627e646e446fc48d9b76f1d5e9da9d12b4395
Reviewed-on: https://review.whamcloud.com/39742
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: James Simmons <jsimmons@infradead.org>
Reviewed-by: Shaun Tancheff <shaun.tancheff@hpe.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Mr NeilBrown [Thu, 27 Aug 2020 03:28:42 +0000 (13:28 +1000)]
LU-6142 mdc: minor function cleanups.
mdc_getxattr_pack() doesn't exist, so remove declaration
mdc_get_lustre_md(), mdc_free_lustre_md(), and
mdc_clear_open_replay_data() are local to mdc_request.c so make them
static and remove declaration from mdc_interal.h.
Test-Parameters: trivial
Signed-off-by: Mr NeilBrown <neilb@suse.de>
Change-Id: Ie925585f51f280a4ad4e3ae7b488dc4166fdaa8d
Reviewed-on: https://review.whamcloud.com/39738
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: James Simmons <jsimmons@infradead.org>
Reviewed-by: Shaun Tancheff <shaun.tancheff@hpe.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Mr NeilBrown [Thu, 27 Aug 2020 01:36:06 +0000 (11:36 +1000)]
LU-6142 osc: minor function cleanups.
osc_cleanup() doesn't exist, so remove declaration.
Return value of osc_extent_release() is always zero and never used, so
change it to 'void'.
Test-Parameters: trivial
Signed-off-by: Mr NeilBrown <neilb@suse.de>
Change-Id: I048e4ec1ec62d2b576f2a1e3c6148c0121cd91b6
Reviewed-on: https://review.whamcloud.com/39737
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: James Simmons <jsimmons@infradead.org>
Reviewed-by: Shaun Tancheff <shaun.tancheff@hpe.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Mr NeilBrown [Thu, 16 Jul 2020 04:56:38 +0000 (14:56 +1000)]
LU-6142 lustre: change various operations structs to const
Nearly all of
struct cl_io_operations
struct cl_lock_operations
struct iam_operations
struct iam_leaf_operations
struct llog_operations
struct lquota_entry_operations
struct md_dir_operations
are now const. The one exception is changelog_orig_logops.
Test-Parameters: trivial
Signed-off-by: Mr NeilBrown <neilb@suse.de>
Change-Id: I91dc9528b62a77f6bcedc80d6066a74a8eeaa3fa
Reviewed-on: https://review.whamcloud.com/39400
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: James Simmons <jsimmons@infradead.org>
Reviewed-by: Arshad Hussain <arshad.hussain@aeoncomputing.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Mr NeilBrown [Thu, 16 Jul 2020 04:07:56 +0000 (14:07 +1000)]
LU-6142 lfsck: make all 'struct lfsck_operations' to const
Operations structures should be 'const' wherever possible.
So change both lfsck_operations and lfsck_assistant_operations
to const.
Test-Parameters: trivial
Signed-off-by: Mr NeilBrown <neilb@suse.de>
Change-Id: Ic457c0e52ceafab58cb788db383957ed4ab39df4
Reviewed-on: https://review.whamcloud.com/39397
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: James Simmons <jsimmons@infradead.org>
Reviewed-by: Lai Siyao <lai.siyao@whamcloud.com>
Reviewed-by: Arshad Hussain <arshad.hussain@aeoncomputing.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Mr NeilBrown [Thu, 16 Jul 2020 03:59:44 +0000 (13:59 +1000)]
LU-6142 lustre: change all 'struct seq_operations' to const
All 'struct seq_operations' are now 'static const'.
Test-Parameters: trivial
Signed-off-by: Mr NeilBrown <neilb@suse.de>
Change-Id: Ia0f6a2863835ec02fb7d283a099eea489f65ab3a
Reviewed-on: https://review.whamcloud.com/39396
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: James Simmons <jsimmons@infradead.org>
Reviewed-by: Arshad Hussain <arshad.hussain@aeoncomputing.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Mr NeilBrown [Thu, 16 Jul 2020 03:49:36 +0000 (13:49 +1000)]
LU-6142 lustre: change super/file/inode operations to const
All 'struct file_operations', 'struct inode_operations', 'struct
export_operations' and 'struct super_operations' are changed to
'const'. This potenetially allows them to be placed in read-only
memory, and ensure they are never changed.
Test-Parameters: trivial
Signed-off-by: Mr NeilBrown <neilb@suse.de>
Change-Id: I8b236f0248eca11f91f11da02fe18be3f6d2e17c
Reviewed-on: https://review.whamcloud.com/39394
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: James Simmons <jsimmons@infradead.org>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Mr NeilBrown [Thu, 16 Jul 2020 00:32:02 +0000 (10:32 +1000)]
LU-6142 obdclass: use cl_object_for_each more broadly.
There are several loops which follow the pattern of the macros
cl_object_for_each() or cl_object_for_each_reverse(), but
don't use those macros.
Change to use the macros. In some case we change the name of an 'obj'
arg to a function to 'top', to avoid overloading the name 'obj'.
Also include some minor formatting improvements.
Signed-off-by: Mr NeilBrown <neilb@suse.de>
Change-Id: I544d4148719a9284f33fd61730bd733c0d40e97c
Reviewed-on: https://review.whamcloud.com/39391
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Wang Shilong <wshilong@whamcloud.com>
Reviewed-by: James Simmons <jsimmons@infradead.org>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Mr NeilBrown [Wed, 15 Jul 2020 06:32:47 +0000 (16:32 +1000)]
LU-6142 lustre: remove module_vars arg to class_register_type()
The module_vars arg to class_register_type() is always NULL. So it
can be removed.
Test-Parameters: trivial
Signed-off-by: Mr NeilBrown <neilb@suse.de>
Change-Id: Ie8d7e79075ba068dec606cc9dfcc38a90e371e5b
Reviewed-on: https://review.whamcloud.com/39383
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Shaun Tancheff <shaun.tancheff@hpe.com>
Reviewed-by: James Simmons <jsimmons@infradead.org>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Mr NeilBrown [Wed, 15 Jul 2020 05:42:23 +0000 (15:42 +1000)]
LU-6142 lov: style cleanups in lov_set_osc_active()
1/ don't pre-declare the function as it is not used before it
is defined.
2/ " TEST ? 1 : 0" is identical to "TEST", so remove the "? 1 : 0".
3/ When the 'then' part of an 'if' ends with a GOTO, there is
not point having an 'else', particularly if there is also
code in the block following the "if". The 'else' code and
the following code should be together.
4/ other minor changes, and conversion of spaces to tabs.
Test-Parameters: trivial
Signed-off-by: Mr NeilBrown <neilb@suse.de>
Change-Id: I00cb8b1efe25c0ad9adc76ba21cca003aed203e2
Reviewed-on: https://review.whamcloud.com/39382
Tested-by: jenkins <devops@whamcloud.com>
Reviewed-by: Arshad Hussain <arshad.hussain@aeoncomputing.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: James Simmons <jsimmons@infradead.org>
Reviewed-by: Lai Siyao <lai.siyao@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Mr NeilBrown [Wed, 15 Jul 2020 06:16:35 +0000 (16:16 +1000)]
LU-6142 lustre: change obd_ioctl_getdata() args
Instead of passing a pointer to 'char *' to obd_ioctl_getdata(),
pass a pointer to 'struct obd_ioctl_data *'.
This seems more natural, and avoid needing extra 'char *data'
variables wherever the function is called.
Test-Parameters: trivial
Signed-off-by: Mr NeilBrown <neilb@suse.de>
Change-Id: I0287cd2fafd071a2580d48a52498469ceded3e79
Reviewed-on: https://review.whamcloud.com/39381
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-by: Shaun Tancheff <shaun.tancheff@hpe.com>
Reviewed-by: James Simmons <jsimmons@infradead.org>
Mr NeilBrown [Wed, 15 Jul 2020 04:57:28 +0000 (14:57 +1000)]
LU-6142 lov: chnage lsm_op_find() to a non-inline function.
Making lsm_op_find() non-inline improves encapsulation and means that
various _ops structures can now be static.
Test-Parameters: trivial
Signed-off-by: Mr NeilBrown <neilb@suse.de>
Change-Id: Ifee2ab4f023e56bfa1e663a0106590cf7858856b
Reviewed-on: https://review.whamcloud.com/39377
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: James Simmons <jsimmons@infradead.org>
Reviewed-by: Shaun Tancheff <shaun.tancheff@hpe.com>
Reviewed-by: Lai Siyao <lai.siyao@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Mr NeilBrown [Thu, 27 Aug 2020 04:06:40 +0000 (14:06 +1000)]
LU-6142 lustre: make various 'struct file_operations' static
These 'struct file_operations' are only used locally, so make them
static.
Except lprocfs_evict_client_fops() which isn't used at all and doesn't
exist, so discard the declaration.
Test-Parameters: trivial
Signed-off-by: Mr NeilBrown <neilb@suse.de>
Change-Id: Ib6c51683c1e765db202b3f72d2accebe17191303
Reviewed-on: https://review.whamcloud.com/39741
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: James Simmons <jsimmons@infradead.org>
Reviewed-by: Shaun Tancheff <shaun.tancheff@hpe.com>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Mr. NeilBrown [Thu, 13 Aug 2020 12:44:25 +0000 (08:44 -0400)]
LU-9859 libcfs: discard TCD_MAX_TYPES
As well as CFS_TCD_TYPE_CNT we have TCD_MAX_TYPES which has a larger
value but a similar meaning. Discard it and just use
CFS_TCD_TYPE_CNT.
Two places relied on the fact that TCD_MAX_TYPES was larger and so
there would be NULLs at the end of the array. Change
them to check the array size properly.
Linux-commit:
22627ffa5e7fae0b0b62f90a8f30fc6dae0fff3c
Change-Id: I73a54a409674d3b569bcbaf72e2296209542b74c
Signed-off-by: Mr. NeilBrown <neilb@suse.de>
Reviewed-on: https://review.whamcloud.com/39668
Reviewed-by: Jian Yu <yujian@whamcloud.com>
Reviewed-by: Yang Sheng <ys@whamcloud.com>
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Mr NeilBrown [Fri, 24 Jan 2020 01:10:55 +0000 (12:10 +1100)]
LU-10391 lnet: allow lnet_connect() to use IPv6 addresses.
Now that we can created and bind IPv6 sockets, connecting to them is
trivial.
Signed-off-by: Mr NeilBrown <neilb@suse.de>
Change-Id: I6032ea8a5937711922d9c318f72bc64c7020cc6d
Reviewed-on: https://review.whamcloud.com/37706
Reviewed-by: Serguei Smirnov <ssmirnov@whamcloud.com>
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Aurelien Degremont <degremoa@amazon.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Mr NeilBrown [Wed, 22 Jan 2020 23:26:37 +0000 (10:26 +1100)]
LU-10391 lnet: allow creation of IPv6 socket.
With this patch, lnet_sock_create() can make IPv6 sockets. If an
interface and destination is given, completely different code is
needed to bind the local address.
When no interface or destination is given, an IPv6 socket is always
created, so lnet_acceptor() needs to request that IPv4 connections are
accepted as well, and lnet_sock_getaddr() needs to detect mapped v4
addresses and present them as true v4 addresses.
Signed-off-by: Mr NeilBrown <neilb@suse.de>
Change-Id: Iaef1c08a9afd578b60f54cc83fd2d6882a021e6b
Reviewed-on: https://review.whamcloud.com/37705
Reviewed-by: Shaun Tancheff <shaun.tancheff@hpe.com>
Reviewed-by: Serguei Smirnov <ssmirnov@whamcloud.com>
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Mr NeilBrown [Sun, 7 Jun 2020 23:24:24 +0000 (19:24 -0400)]
LU-9679 ptlrpc: list_for_each improvements.
1/ use list_for_each_entry_safe() instead of
list_for_each_safe() and similar.
2/ use list_first_entry() and list_last_entry() where appropriate.
3/ When removing everything from a list, use
while ((x = list_first_entry_or_null()) {
as it makes the intent clear
Linux-commit:
ef8e5dbbb09035a0c41aa47a328e6248702d4d2b
Signed-off-by: Mr NeilBrown <neilb@suse.de>
Change-Id: I490afcb70d18170e4ecc5d062d327d141668d3db
Reviewed-on: https://review.whamcloud.com/39133
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: James Simmons <jsimmons@infradead.org>
Reviewed-by: Shaun Tancheff <shaun.tancheff@hpe.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Elena Gryaznova [Tue, 22 Dec 2020 17:41:22 +0000 (20:41 +0300)]
LU-14275 tests: add ior_CLEANUP
Sometimes we need to run read IOR test separately from
write test. IOR -r test requires file exists before test starts.
If ior_CLEANUP set to false $testdir is not removed and
$testdir/iorData file can be used for IOR -r further test.
Test-Parameters: trivial testlist=parallel-scale
Signed-off-by: Elena Gryaznova <elena.gryaznova@hpe.com>
HPE-bug-id: LUS-9560
Reviewed-by: Vladimir Saveliev <vladimir.saveliev@hpe.com>
Reviewed-by: Sergey Cheremencev <sergey.cheremencev@hpe.com>
Change-Id: Iac790976a46db2bf5be374a42e3facdce074228b
Reviewed-on: https://review.whamcloud.com/41079
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Andreas Dilger [Sat, 30 Jan 2021 19:42:36 +0000 (12:42 -0700)]
LU-14388 utils: always enable ldiskfs project quota
Always enable project quota for newly-formatted ldiskfs filesystems.
Signed-off-by: Andreas Dilger <adilger@whamcloud.com>
Change-Id: I1b0f745bc04b5c42592bcc4fd9823d068fef2a79
Reviewed-on: https://review.whamcloud.com/41370
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Wang Shilong <wshilong@whamcloud.com>
Reviewed-by: Li Xi <lixi@ddn.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Mr NeilBrown [Fri, 30 Oct 2020 04:18:21 +0000 (15:18 +1100)]
LU-14353 obd: move debug.c to obdecho
The code in obdclass/debug.c is only used from obdecho/, so move the
file there and clean it up a little on the way (e.g. spaces to TABs).
Similarly lustre_debug.h contains very little that is used outside of
obdecho. Move that little (LL_CDEBUG_PAGE()) into lustre_lib.h, and
add the declarations needed for debug.c into echo_internal.h.
Test-Parameters: trivial
Signed-off-by: Mr NeilBrown <neilb@suse.de>
Change-Id: Ia1bcf3f3e1453130ed1adc75918bdad23eb5b5c7
Reviewed-on: https://review.whamcloud.com/41290
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>
Andreas Dilger [Wed, 6 Jan 2021 01:46:03 +0000 (18:46 -0700)]
LU-14305 ldiskfs: add parameters for mb_c123_threshold
Add mount options for /sys/fs/ldiskfs/*/mb_c[123]_threshold values
so that they can be set persistently via mount options.
The /sys/fs/ldiskfs/*/mb_c[123]_threshold values are always shown
rounded down to the next lower percentage value due to integer
division, since internal values are stored as blocks for efficiency.
Round up the values shown to the next percent to match what was
used to originally set these parameters.
Fixes:
95f8ae567749 ("LU-12103 ldiskfs: don't search large block range if disk full")
Signed-off-by: Andreas Dilger <adilger@whamcloud.com>
Change-Id: Ie36a6667f8bca7481aa8179ab5b97c85d449d619
Reviewed-on: https://review.whamcloud.com/41193
Reviewed-by: Artem Blagodarenko <artem.blagodarenko@hpe.com>
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Jian Yu <yujian@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Mr NeilBrown [Mon, 23 Nov 2020 02:26:22 +0000 (13:26 +1100)]
LU-14289 ptlrpc: move heap.c from libcfs to ptlrpc
cfs_binheap is only used by the nrs code in ptlrpc, so move it out of
libcfs to the one place it is used.
Test-Parameters: trivial
Signed-off-by: Mr NeilBrown <neilb@suse.de>
Change-Id: I4687d4de0a83e0ce106d4f4cfd5ea95faef27dfc
Reviewed-on: https://review.whamcloud.com/41129
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>
Mr NeilBrown [Thu, 29 Oct 2020 04:01:35 +0000 (15:01 +1100)]
LU-14291 lustre: only include nrs headers when needed
Various lustre_nrs_*.h header are only needed on the server. Make
this explicit, so it is easier to see which files are needed on the
client.
Test-Parameters: trivial
Signed-off-by: Mr NeilBrown <neilb@suse.de>
Change-Id: I9bc30acc4f62ea9af516055edbb7833919811089
Reviewed-on: https://review.whamcloud.com/41128
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: James Simmons <jsimmons@infradead.org>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Mr NeilBrown [Thu, 29 Oct 2020 03:41:18 +0000 (14:41 +1100)]
LU-14291 ptlrpc: support nrs_delay for client-only builds
nrs_delay is intended for use on the client, but is currently disabled
for client-only builds. So move the ptlrpc_nrs_policy_register() call
outside of the #ifdef.
Also change the nrs modules that *are* server-only to not have #ifdef
in the C file, but that rather to be identified in the Makefile as
only being built when server code is wanted.
Test-Parameters: trivial
Signed-off-by: Mr NeilBrown <neilb@suse.de>
Change-Id: Ie25ed5ac837ff5f891824fad31ab258923840ec0
Reviewed-on: https://review.whamcloud.com/41127
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>
Arshad Hussain [Thu, 31 Dec 2020 19:12:05 +0000 (00:42 +0530)]
LU-14285 utils: Add error message when osd_init fails
It was observed that mkfs.lustre fails silently under
function osd_init(). One such case where it silently
fails is when back-end is 'zfs' and libz.so library
is not loaded or under search path.
This patch adds error message in case osd_init() call
fails. It makes end user observe the failure more easily.
Before Patch:
~~~~~~~~~~~~~
$ echo $LD_LIBRARY_PATH
/root/zfs_spl_install/0840/zfs/lib
$ unset LD_LIBRARY_PATH
$ mkfs.lustre --reformat --index 0 --backfstype=zfs --fsname=lustre --mgsnode=192.168.50.140@tcp --mgs --mdt gpool/metadata /dev/sdb1
$ echo $?
22
After Patch:
~~~~~~~~~~~~
$ echo $LD_LIBRARY_PATH
/root/zfs_spl_install/0840/zfs/lib
$ unset LD_LIBRARY_PATH
$ mkfs.lustre --reformat --index 0 --backfstype=zfs --fsname=lustre --mgsnode=192.168.50.140@tcp --mgs --mdt gpool/metadata /dev/sdb1
mkfs.lustre: osd_init() failed : 22 (Invalid argument)
$
Test-Parameters: trivial
Signed-off-by: Arshad Hussain <arshad.hussain@aeoncomputing.com>
Change-Id: Ie75b00974852d7a4b95b52d26501e55b66db5891
Reviewed-on: https://review.whamcloud.com/41118
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>
Andreas Dilger [Thu, 24 Dec 2020 01:30:46 +0000 (18:30 -0700)]
LU-12477 llite: remove unused ll_teardown_mmaps()
The ll_teardown_mmaps() function is no longer used and can be
removed.
Fixes:
f114d4dac6f8 ("LU-12477 llite: use generic_error_remove_page()")
Test-Parameters: trivial
Signed-off-by: Andreas Dilger <adilger@whamcloud.com>
Change-Id: I42506210e028da3bf8036a28d4d61b1e7d3ebbe5
Reviewed-on: https://review.whamcloud.com/41086
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: James Simmons <jsimmons@infradead.org>
Reviewed-by: Neil Brown <neilb@suse.de>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
James Nunez [Tue, 22 Dec 2020 15:56:45 +0000 (08:56 -0700)]
LU-930 doc: fix format man page sections for lctl
For lctl man pages, add proper formatting for the NAME
and TITLE sections so that 'man -k lctl*' reports the
real explanation for the command/utility and not
"(unknown subject)".
Modifications in this patch include:
- add a dash between multi-word name commands in NAME
- format the date to YYYY-MM-DD in TITLE
- capitalize command in TITLE
- change hyphen (-) to underscore in command where necessary
Test-Parameters: trivial
Signed-off-by: James Nunez <jnunez@whamcloud.com>
Change-Id: I946a3683b0a48b662c4e3b7b5783d389fa710e08
Reviewed-on: https://review.whamcloud.com/41075
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: John L. Hammond <jhammond@whamcloud.com>
Elena Gryaznova [Tue, 22 Dec 2020 15:00:08 +0000 (18:00 +0300)]
LU-14272 tests: different mpirun options for different users
Patch adds the ability to use different mpirun options for
different mpi users, for example we use:
--wdir=/home/quota15_1 for user quota15_1
and
--wdir=/home/quota15_2 for user quota15_2.
Per mpi user option can be specified after ':' in ha_mpi_users:
Example:
ha_mpi_users="quota15_1:--wdir=/home/quota15_1
quota15_2:--wdir=/home/quota15_2
quota15_3
quota15_4:--wdir=/home/quota15_4"
Several options for user can be set as:
ha_mpi_users="user1:<option11>
user1:<option12>
user2:<option21>"
MPIRUN_OPTIONS will be applied for all users.
Patch adds the ability to execute the command before loads start.
Example: ha_precmd="for i in 1 2 3 4; do
lfs quota -v -u quota15_$i $ha_test_dir; done"
Test-Parameters: trivial
Signed-off-by: Elena Gryaznova <elena.gryaznova@hpe.com>
HPE-bug-id: LUS-9069
Reviewed-by: Vladimir Saveliev <vladimir.saveliev@hpe.com>
Reviewed-by: Sergey Cheremencev <sergey.cheremencev@hpe.com>
Change-Id: I35feadcc856dbc3d6d6d5f2720ab49c6c111161b
Reviewed-on: https://review.whamcloud.com/41072
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>
Elena Gryaznova [Tue, 22 Dec 2020 14:06:09 +0000 (17:06 +0300)]
LU-14271 tests: add new node crash method
Patch adds a new node crash method: node crash
is simulated by command executed on the victim nodes.
Example:
POWER_DOWN=sysrqcrash
triggers a crash on all victim nodes.
Test-Parameters: trivial
Signed-off-by: Elena Gryaznova <elena.gryaznova@hpe.com>
HPE-bug-id: LUS-8162
Reviewed-by: Vladimir Saveliev <vladimir.saveliev@hpe.com>
Reviewed-by: Sergey Cheremencev <sergey.cheremencev@hpe.com>
Tested-by: Vitaly Fertman <vitaly.fertman@hpe.com>
Change-Id: Id39b0f58fbb59236b3af51516fe968c888b04a17
Reviewed-on: https://review.whamcloud.com/41071
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>
Elena Gryaznova [Tue, 22 Dec 2020 13:21:47 +0000 (16:21 +0300)]
LU-14270 tests: delay node's power up
In CS a node in UNCLEAN state can be STONITHed after
it has passed power down and power up already.
The patch adds ha_power_up_delay() in order to delay
node's power up until CRM state became OFFLINE when
failover pair is set and for $NODE_UP_DELAY seconds otherwise.
LOAD_TIMEOUT is added in order to ha_load_timeout be tunable.
The failover pair list for all victims are to be set via new
-f option.
ha_powermanage() is added to control node state by pm, (designed
by Maxim Storchak <maxim.storchak@hpe.com>).
PM_HOST is added in order to specify powermanage host, default
is $(hostname).
Test-Parameters: trivial
Signed-off-by: Elena Gryaznova <elena.gryaznova@hpe.com>
HPE-bug-id: LUS-8610
Reviewed-by: Maxim Storchak <maxim.storchak@hpe.com>
Reviewed-by: Andriy Skulysh <andriy.skulysh@hpe.com>
Change-Id: Ia192200b9ff61c17a67fc16858938f24ade46e23
Reviewed-on: https://review.whamcloud.com/41070
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-by: Wei Liu <sarah@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Alex Zhuravlev [Thu, 17 Dec 2020 09:15:50 +0000 (12:15 +0300)]
LU-14262 utils: lfs to set component flags by pool name
so it'd be easy to set flags (like prefer) on the components
residing on specific OST identified by pool.
Signed-off-by: Alex Zhuravlev <bzzz@whamcloud.com>
Change-Id: I733f92fe186682dc8d34512edf75b49e565c457f
Reviewed-on: https://review.whamcloud.com/41024
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-by: John L. Hammond <jhammond@whamcloud.com>
Tested-by: jenkins <devops@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Mr NeilBrown [Wed, 9 Dec 2020 02:11:18 +0000 (13:11 +1100)]
LU-14195 lustre: remove 'fs' from 'struct lvfs_run_ctxt'
The code protected by push_ctxt() and pop_ctx() never tries to access
any user-space data, so call set_fs() to KERNEL_DS is not needed.
So remove the 'fs' field and related code.
In linux-5.10 this code fails to compile as set_fs() is deprecated.
Test-Parameters: trivial
Signed-off-by: Mr NeilBrown <neilb@suse.de>
Change-Id: Idb2744d656dc4228375b6da54673e38cc1c112f5
Reviewed-on: https://review.whamcloud.com/40910
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: Alex Zhuravlev <bzzz@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Mr NeilBrown [Wed, 9 Dec 2020 02:00:16 +0000 (13:00 +1100)]
LU-14195 osd: don't use set_fs() for ->fiemap() calls.
->fiemap() only accesses kernel-space data, so does not need, and
never has needed, set_fs() calls.
In Linux 5.10, these calls are deprecated.
So remove the unnecessary code.
Test-Parameters: trivial
Signed-off-by: Mr NeilBrown <neilb@suse.de>
Change-Id: Id336855b4787ddbf656dfa3b8d0b12f663564795
Reviewed-on: https://review.whamcloud.com/40909
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: James Simmons <jsimmons@infradead.org>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Mr NeilBrown [Wed, 9 Dec 2020 00:28:16 +0000 (11:28 +1100)]
LU-14195 build: Adjust Makefile for Linux build changes.
Since v5.10-rc1~51^2~19, "KBUILD_BUILTIN" has been unset
for module builds. This means that "targets-for-builtin"
isn't built, and that is how "extra-y" is built.
So we need another way to force LUSTRE_KERNEL_TEST to be built.
Since v5.6-rc1~1^2~5 any target listed in "always-y" will always get
built. So we can assign LUSTRE_KERNEL_TEST to this macro.
Assigning both macros is safe, even for those kernels which include
both in the list of targets.
Test-Parameters: trivial
Signed-off-by: Mr NeilBrown <neilb@suse.de>
Change-Id: I508b3710579c068dec93baf81ee383f3f03bd370
Reviewed-on: https://review.whamcloud.com/40907
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Aurelien Degremont <degremoa@amazon.com>
Reviewed-by: James Simmons <jsimmons@infradead.org>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Andreas Dilger [Thu, 3 Dec 2020 09:40:35 +0000 (02:40 -0700)]
LU-14178 ldlm: return error from ldlm_namespace_new()
Return the underlying error in ldlm_namespace_new() from
ldlm_namespace_sysfs_register() to the caller instead of NULL.
Otherwise, the callers convert the NULL to -ENOMEM and this
is incorrectly reported as an allocation error to the user.
sysfs: cannot create duplicate filename
'/fs/lustre/ldlm/namespaces/lustre-OST0002-osc-
ffff89f33be70000'
mount.lustre: mount mgs:/lfs at /lfs failed: Cannot allocate memory
Change ldlm_namespace_new() to return errors via PTR_ERR() and
change the callers to use IS_ERR().
Fix associated CERROR() messages to follow proper code style.
Signed-off-by: Andreas Dilger <adilger@whamcloud.com>
Change-Id: I7d590c7242399549b32b1c4189e46ff8748c8096
Reviewed-on: https://review.whamcloud.com/40851
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Alex Zhuravlev <bzzz@whamcloud.com>
Reviewed-by: James Simmons <jsimmons@infradead.org>
Reviewed-by: Neil Brown <neilb@suse.de>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Mr NeilBrown [Mon, 26 Oct 2020 02:30:22 +0000 (13:30 +1100)]
LU-14073 ofd: remove use of smp_read_barrier_depends()
Linux 5.9 removes smp_read_barrier_depends(), so lustre must stop
using it.
There is only one use: in ofd_access_log.c.
This use is unnecessary and can simply be removed.
The code is based on "Documentation/core-api/circular-buffers.rst"
which gives no indication that this barrier is needed.
The comment say its purpose is to ensure the index is read before the
data is read. This is unnecessary.
The data is written in osl_write_entry(), then a barrier is issued
(smp_store_release) before the ->head is written.
oal_read_entry() issues a barrier (smp_load_acquire()) before reading
that head.
'tail' is read without a barrer, but it then compared against ->head
in CIRC_CNT(). Even if reading ->tail was racey, the fact that
comparing it wilth ->head succeeded means that the data written at
->tail must have been safely written, and we can now read it without
any further barrier.
Test-Parameters: trivial
Signed-off-by: Mr NeilBrown <neilb@suse.de>
Change-Id: I9d0f0aeb67e1188d2012f4ae2e14b3656211c3e2
Reviewed-on: https://review.whamcloud.com/40394
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: John L. Hammond <jhammond@whamcloud.com>
Reviewed-by: James Simmons <jsimmons@infradead.org>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
John L. Hammond [Tue, 20 Oct 2020 14:20:35 +0000 (09:20 -0500)]
LU-14047 lustre: change EWOULDBLOCK to EAGAIN
On linux, EWOULDBLOCK has always been defined as an alias for
EAGAIN. In the interest of readability we should not use two names for
the same thing. So change the remaining uses of EWOULDBLOCK to EAGAIN
and add EWOULDBLOCK||EAGAIN to spelling.txt.
Signed-off-by: John L. Hammond <jhammond@whamcloud.com>
Change-Id: Ib48b8a1e58bfa961d2a4ba411c038c476bfc300d
Reviewed-on: https://review.whamcloud.com/40307
Tested-by: jenkins <devops@whamcloud.com>
Reviewed-by: Neil Brown <neilb@suse.de>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Yingjin Qian <qian@ddn.com>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Mr NeilBrown [Fri, 16 Oct 2020 06:18:29 +0000 (17:18 +1100)]
LU-13783 libcfs: switch from ->mmap_sem to mmap_lock()
In Linux 5.8, ->mmap_sem is gone and the preferred interface
for locking the mmap is to suite of mmap*lock() functions.
So provide those functions when not available, and use them
as needed in Lustre.
Signed-off-by: Mr NeilBrown <neilb@suse.de>
Change-Id: I4ce3959f9e93eae10a7b7db03e2b0a1525723138
Reviewed-on: https://review.whamcloud.com/40288
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: James Simmons <jsimmons@infradead.org>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
James Simmons [Tue, 3 Nov 2020 14:27:43 +0000 (09:27 -0500)]
LU-9325 osd-ldisk: replace simple_strto* with kstr* functions
The parsing of mount parameters from the config llog is done in
some cases with simple_strto* which is considered obsolete.
Replace simple_strto* with the kstrto* equivalent functions.
Change-Id: I7c26d14d02828c9f9a96f31a086a65bb39f3ea87
Signed-off-by: James Simmons <jsimmons@infradead.org>
Reviewed-on: https://review.whamcloud.com/40119
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Neil Brown <neilb@suse.de>
Reviewed-by: Alex Zhuravlev <bzzz@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
James Simmons [Thu, 10 Sep 2020 14:35:55 +0000 (10:35 -0400)]
LU-9859 lod: use linux kernel bitmap API
Now that modern Linux kernels support a bitmap API we can move
away from the libcfs specific bitmap API. This patch changes
the last bitmap in the lod module to use the Linux kernel
bitmap API.
Change-Id: I92c494bf2af62e31d7b9527b3f44580322e48fd3
Signed-off-by: James Simmons <jsimmons@infradead.org>
Reviewed-on: https://review.whamcloud.com/39876
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Yang Sheng <ys@whamcloud.com>
Reviewed-by: Neil Brown <neilb@suse.de>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Mr. NeilBrown [Thu, 10 Sep 2020 13:49:30 +0000 (09:49 -0400)]
LU-9859 libcfs: replace all CFS_CAP_* macros with CAP_*
Lustre defines a few CFS_CAP_* macros which are exactly the
same as the corresponding CAP_* macro, with one exception.
CFS_CAP_SYS_BOOT is 23
CAP_SYS_BOOT is 22.
CFS_CAP_SYS_BOOT is only used through CFS_CAP_FS_MASK and
causes capability 23 (CAP_SYS_NICE) to be dropped in certain
circumstances.
It is probable that the intention was to drop CAP_SYS_BOOT,
and this is what is now done.
CFS_CAP_CHOWN_MASK and CFS_CAP_SYS_RESOURCE_MASK are never
used, so they have been removed.
Linux-commit:
5ebaa2d14850205e44757c4d5fdd4097712d01ef
Change-Id: Ifb90c0a36e204c76b90ff23ac609345d11b878da
Signed-off-by: Mr. NeilBrown <neilb@suse.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Reviewed-on: https://review.whamcloud.com/39875
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Tested-by: Oleg Drokin <green@whamcloud.com>
Andriy Skulysh [Thu, 24 Oct 2019 12:11:25 +0000 (15:11 +0300)]
LU-13100 lov: grant deadlock if same OSC in two components
The same osc can be involved in several components but osc layer
leaves active last used extent, so an RPC can't be sent if grants
are required from the same OST for another component.
Add cl_io_extent_release() to release active extent before
switching to the next component.
Change-Id: Idadda6eaecd1d47b78880c81e1fb7513d5be2419
Cray-bug-id: LUS-8038
Signed-off-by: Andriy Skulysh <c17819@cray.com>
Reviewed-by: Vitaly Fertman <c17818@cray.com>
Reviewed-by: Alexander Zarochentsev <c17826@cray.com>
Reviewed-by: Andrew Perepechko <c17827@cray.com>
Reviewed-on: https://review.whamcloud.com/37095
Tested-by: jenkins <devops@whamcloud.com>
Reviewed-by: Andrew Perepechko <andrew.perepechko@hpe.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Bobi Jam <bobijam@hotmail.com>
Reviewed-by: Mike Pershin <mpershin@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Rahul Deshmkuh [Thu, 14 Jul 2016 06:02:45 +0000 (23:02 -0700)]
LU-7853 lod: fixes bitfield in lod qos code
Updating bitfields in struct lod_qos struct is protected
by lq_rw_sem in most places but an update can be lost
due unprotected bitfield access from
lod_qos_thresholdrr_seq_write() and qos_prio_free_store().
This patch fixes it by replacing bitfields with named bits
and atomic bitops.
Cray-bug-id: LUS-4651
Signed-off-by: Rahul Deshmukh <rahul.deshmukh@seagate.com>
Signed-off-by: Alexander Zarochentsev <c17826@cray.com>
Change-Id: I28299ce4960e91be551d7f6e43a3b598daf4d7a2
Reviewed-on: https://review.whamcloud.com/18812
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Alexander Zarochentsev <alexander.zarochentsev@hpe.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Andreas Dilger [Tue, 26 Jan 2021 02:14:10 +0000 (18:14 -0800)]
LU-14180 utils: verify setstripe comp_end is valid
Verify that the "lfs setstripe -E <component_end>" value is valid.
Otherwise, if "-S" is not specified at the same time, then an
invalid file layout can be created and the file cannot be deleted
normally, only via "lfs rmdif <FID>".
Allow values < 4096 (e.g. '64' or '128' which would all be invalid
anyway) to be interpreted as KiB units.
Update usage messages and man pages to match.
Signed-off-by: Andreas Dilger <adilger@whamcloud.com>
Signed-off-by: Jian Yu <yujian@whamcloud.com>
Change-Id: I47fe7729ffd447c1c1cc098e5117e456263ebbe5
Reviewed-on: https://review.whamcloud.com/41239
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>
Artem Blagodarenko [Wed, 14 Oct 2020 01:54:36 +0000 (21:54 -0400)]
LU-14207 mgs: delete "add failnid" sections on replace_nids
Replace_nids left old nids in add_conn field of failnid
section of client llog. This leads to connecton errors.
Let's delete such sections. New failnids, if any, are
added by replace_nids.
Change-Id: I7fab00827035bd864aeb95fb4852a59c458bb2ba
HPE-bug-id: LUS-9440
Signed-off-by: Artem Blagodarenko <artem.blagodarenko@hpe.com>
Reviewed-by: Vladimir Saveliev <c17830@cray.com>
Reviewed-by: Alexander Boyko <c17825@cray.com>
Tested-by: Jenkins Build User <nssreleng@cray.com>
Reviewed-on: https://review.whamcloud.com/40930
Tested-by: jenkins <devops@whamcloud.com>
Reviewed-by: Alexander Boyko <alexander.boyko@hpe.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Alexander Boyko [Tue, 24 Nov 2020 05:34:11 +0000 (00:34 -0500)]
LU-13974 llog: check stale osp object
The logic of osp_attr_get has 2 path,
1) return attributes from a cache for health osp object
2) make an out update request and return attributes for stale
osp object, object lose stale state.
When some out update request with llog writes failed, osp object
become stale. But llog handle stay inconsistent (bitmap,count,
last_index), and a next llog_add->llog_osd_write_rec do dt_attr_get,
gets attributes and makes osp object valid, and uses wrong llog
handle data. The result is index jump at llog file - recX, recX+2.
And it makes an error during update log processing if failover take
a place.
The fix adds dt_object_stale function to check osp_object.
llog_osd_write_rec check it and return ESTALE. llog_add would fail
with ESTALE error and doesn't corrupt update log.
HPE-bug-id: LUS-9030
Signed-off-by: Alexander Boyko <alexander.boyko@hpe.com>
Change-Id: Iadf53fd816e1c5bde0a19d4c537f0408796c864a
Reviewed-on: https://review.whamcloud.com/40742
Reviewed-by: Alexander Zarochentsev <alexander.zarochentsev@hpe.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>
Alex Zhuravlev [Tue, 12 Nov 2019 17:45:50 +0000 (20:45 +0300)]
LU-12961 mdd: avoid double call to mdd_changelog_fini()
the first call is done from mdd_prepare() as part of error
handling, another call is done from mdd_device_shutdown().
in the similar cases fini routines checks whether the state
is initialized. e.g. mdd_orphan_index_fini() releases the
object and sets mdd_orphans to NULL, then all subsequent
calls to mdd_orphan_index_fini() return immediately.
mdd_changelog_fini() can't do this way, so the excplicit
state has been introduced.
Change-Id: Ifd21569e68c836f44bb59adea4e8fed6ccef1c7b
Signed-off-by: Alex Zhuravlev <bzzz@whamcloud.com>
Reviewed-on: https://review.whamcloud.com/36737
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-by: John L. Hammond <jhammond@whamcloud.com>
Andreas Dilger [Wed, 17 Feb 2021 23:50:59 +0000 (16:50 -0700)]
LU-14439 build: require a newer version of e2fsprogs
Require a build version of e2fsprogs-1.44.3 for osd-ldiskfs builds to
get EXT2_FLAG_IGNORE_SB_ERRORS and ext2fs_has_feature_{mmp,quota}()
functions.
Require a minimum install version of ldiskfsprogs-1.45.6.wc1.
Test-Parameters: trivial
Fixes:
7dc8aa7e7848 ("LU-13241 utils: use libext2fs for ldiskfs operations")
Signed-off-by: Andreas Dilger <adilger@whamcloud.com>
Change-Id: I240823266c7342338c72126011485120543ebbe5
Reviewed-on: https://review.whamcloud.com/41682
Reviewed-by: Olaf Faaland-LLNL <faaland1@llnl.gov>
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: James Simmons <jsimmons@infradead.org>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
Stephane Thiell [Thu, 11 Feb 2021 00:15:02 +0000 (16:15 -0800)]
LU-13609 mgs: fix config_log buffer handling
Fix buffer handling in mgs_list_logs() to list all MGS config_logs
using multiple ioctl calls when we have a large number of targets.
Fixes:
1d97a8b4cd3d ("LU-13609 llog: list all the log files correctly on MGS/MDT")
Signed-off-by: Stephane Thiell <sthiell@stanford.edu>
Change-Id: I1bf32e918e242f4da83c3d1624b7285a18a88d01
Reviewed-on: https://review.whamcloud.com/41478
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-by: Emoly Liu <emoly@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
git://git.whamcloud.com / fs / lustre-release.git / log