Whamcloud - gitweb
fs/lustre-release.git
21 years agoUse PTL_MD_MAX_PAGES instead of PTL_MD_MAX_IOV where needed, so we send the
adilger [Fri, 14 Nov 2003 19:27:01 +0000 (19:27 +0000)]
Use PTL_MD_MAX_PAGES instead of PTL_MD_MAX_IOV where needed, so we send the
right number of pages over the wire for ia64/i386 interoperability.
b=1821

21 years agoliblustre: update lrun script.
ericm [Fri, 14 Nov 2003 15:05:48 +0000 (15:05 +0000)]
liblustre: update lrun script.

21 years agoliblustre:
ericm [Fri, 14 Nov 2003 11:13:26 +0000 (11:13 +0000)]
liblustre:
  add more test items into replay_single.

21 years agoliblustre: add more test in replay_single, and some other adjustment.
ericm [Fri, 14 Nov 2003 09:37:33 +0000 (09:37 +0000)]
liblustre: add more test in replay_single, and some other adjustment.

21 years agoI touched one thing in lustre_mds.h, and I was dismayed to see the
phil [Fri, 14 Nov 2003 09:24:31 +0000 (09:24 +0000)]
I touched one thing in lustre_mds.h, and I was dismayed to see the
entire source tree rebuild!

It took just a few minutes to remove a number of #includes which
violated all manner of abstraction boundary.  ericm assures me that I
have not broken anything major in liblustre.

Also, I moved a good chunk of inline functions into llite_internal.h
(which is where I secretly believe that most of lustre_lite.h will end up),
and renamed ll_ino2fid (which no longer takes an inode) to mdc_pack_fid.

21 years ago- fix del_route
rread [Fri, 14 Nov 2003 01:12:31 +0000 (01:12 +0000)]
- fix del_route

21 years ago- update this patch to apply against vanilla-2.4.22
alex [Thu, 13 Nov 2003 20:49:31 +0000 (20:49 +0000)]
- update this patch to apply against vanilla-2.4.22

21 years agoSilence most of the warnings in the x86_64 build, and kill some Real Bugs in
zab [Thu, 13 Nov 2003 20:21:48 +0000 (20:21 +0000)]
Silence most of the warnings in the x86_64 build, and kill some Real Bugs in
the process.  (sanity checked in an x86 uml build, as well)

- include linux/init.h from some modules that use __init and company
- x86_64 has 64 bit longs but uses 'long long' for u64 in the kernel.  we
  add its own specific section to LP*64 definitions.  Sadly, there are still
  a huge number of warnings from this gcc when we try to LP*64 print uint64_t
  obdo arguments.
- use min_t in some defines to avoid the duplicate const prefacing that comes
  with nesting min()
- LPX64 printing -1UL doesn't work so well
- i_nlink can be any of unsigned short, int, or long.  when printing cast
  to unsigned long.
- kern_addr_valid takes a 'unsigned long' argument and shifts it.. sigh.
- don't #define EXPORT_SYMTAB in the source, that's the job of the kernel
  makefiles
- x86_64/suse has THREAD_SIZE defined in userspace
- s/long flags/unsigned long flags/.  Again.
- va_list ap = NULL; is not legal, as far as I know.  it doesn't seem
  to actually do anything in this code path either.
- x86_64 passes CONFIG_X86 but had thread->rsp as the stack pointer.  this
  code is going away eventually anyway.
- print a pointer with %p, not %x
- some portals callback pointer definitions used size_t while some nal's
  functions used ptl_size_t in the function defintions.  I switched the
  functions over to avoid pushing changes on all the nals.

21 years ago- ide_end_request() was used wrong way: fixed
alex [Thu, 13 Nov 2003 15:30:09 +0000 (15:30 +0000)]
- ide_end_request() was used wrong way: fixed

NOTE: tested on the all related series

21 years ago- kmem_cache_validate patch has been removed from all the series
alex [Thu, 13 Nov 2003 15:06:15 +0000 (15:06 +0000)]
- kmem_cache_validate patch has been removed from all the series

NOTE: all the supported series still build

21 years agoliblustre: zconf adjustment for recovery_small and replay_single.
ericm [Thu, 13 Nov 2003 13:01:48 +0000 (13:01 +0000)]
liblustre: zconf adjustment for recovery_small and replay_single.

21 years agoliblustre:
ericm [Thu, 13 Nov 2003 06:56:20 +0000 (06:56 +0000)]
liblustre:
 - adjust the way of zeroconf
 - bring lltest/recovery_small to zeroconf

21 years ago b:2215 - OSTs fetch unlink llog records from MDS post replay
tianying [Thu, 13 Nov 2003 05:57:06 +0000 (05:57 +0000)]
  b:2215 - OSTs fetch unlink llog records from MDS post replay

     1. add lop_connect and lop_precleanup to llog_operations
     2. rename llog_obd_ctxt to llog_ctxt; llog_commit_data to llog_canceld_ctxt
     3. split out llog functions in llog_client.c and llog_server.c and remove llogd.c
     4. add one test-59 to sanity.sh to verify cancellation of llog records async
     5. fix calling of mds_cleanup_orphans, add test-34 to replay-single.sh
     6. fix some codes about llog

21 years ago- return replay-ost-single to runable state
rread [Thu, 13 Nov 2003 01:27:36 +0000 (01:27 +0000)]
- return replay-ost-single to runable state
  now just need to pass

- commonize the zconf mount/umount

21 years agoFix ia64 compile warnings. Please folks, use LPU64 instead of %llu.
adilger [Thu, 13 Nov 2003 01:00:08 +0000 (01:00 +0000)]
Fix ia64 compile warnings.  Please folks, use LPU64 instead of %llu.

21 years ago- enable KMOD
rread [Wed, 12 Nov 2003 23:17:22 +0000 (23:17 +0000)]
- enable KMOD

(CONFIG_MODE_TT=y is added by make oldconfig)

21 years agosample modules.conf for zeroconf
rread [Wed, 12 Nov 2003 23:14:01 +0000 (23:14 +0000)]
sample modules.conf for zeroconf

21 years agoRemove kmem_cache_validate_hp from hp-pnnl-2.4.20 series.
adilger [Wed, 12 Nov 2003 23:11:46 +0000 (23:11 +0000)]
Remove kmem_cache_validate_hp from hp-pnnl-2.4.20 series.

21 years agob=2244
rread [Wed, 12 Nov 2003 21:16:28 +0000 (21:16 +0000)]
b=2244
r=shaver

Pass force flag to ldlm_cleanup.

21 years ago- fix set_route
rread [Wed, 12 Nov 2003 21:11:24 +0000 (21:11 +0000)]
- fix set_route

- add kptlrouter to list of modules

21 years agoAdd process pid into console log messages.
adilger [Wed, 12 Nov 2003 19:49:42 +0000 (19:49 +0000)]
Add process pid into console log messages.

21 years agoliblustre: support zeroconf.
ericm [Wed, 12 Nov 2003 14:15:15 +0000 (14:15 +0000)]
liblustre: support zeroconf.
but current existed tests still use dumpfile, switch to zconf later.

21 years agob=2134
wangchao [Wed, 12 Nov 2003 07:56:04 +0000 (07:56 +0000)]
b=2134
r=shaver
add a test case of unix domain socket test to sanity.sh (test case 54)

21 years agoliblustre: add missing lib for shared so creation.
ericm [Wed, 12 Nov 2003 07:31:46 +0000 (07:31 +0000)]
liblustre: add missing lib for shared so creation.

21 years agoDon't have a separate (inconsistent) MTU define for portals routers.
adilger [Wed, 12 Nov 2003 06:27:19 +0000 (06:27 +0000)]
Don't have a separate (inconsistent) MTU define for portals routers.
Print out portals router messages with nicely formatted NIDs.

21 years ago- add /sbin/mount.lustre to rpm
rread [Wed, 12 Nov 2003 06:18:48 +0000 (06:18 +0000)]
- add /sbin/mount.lustre to rpm

21 years agocleanup replay-single.sh
rread [Wed, 12 Nov 2003 06:18:10 +0000 (06:18 +0000)]
cleanup replay-single.sh

21 years agob=2243
rread [Wed, 12 Nov 2003 02:41:21 +0000 (02:41 +0000)]
b=2243
fix router config breakage.  Now that the router config commands are
NAL_CMDs, need to add a gw_nal to pcfg.

21 years agocreate ibnal/Makefile
rread [Wed, 12 Nov 2003 00:54:34 +0000 (00:54 +0000)]
create ibnal/Makefile

21 years agoQuiet warnings.
shaver [Wed, 12 Nov 2003 00:17:01 +0000 (00:17 +0000)]
Quiet warnings.

21 years agoAdd new "lustre" fs type which supports only zeroconf mounts.
rread [Tue, 11 Nov 2003 23:13:14 +0000 (23:13 +0000)]
Add new "lustre" fs type which supports only zeroconf mounts.
 - old zeroconf client code removed from lustre-lite, and the lconf
   --zeroconf option deleted
 - common code factored out for ll/lustre fill_super and put_super
 - lconf still uses lustre_lite (but not for long)

llmount will be used by mount for lustre filesystems if copied to
/sbin/mount.lustre:

  mount -t lustre mds_host:/mds_servicee/profile  /mnt/lustre

Multiple mounts of the same filesystem are supported.

Remove unused mds and filter nspath code.

21 years ago* fixed kportal_router_cmd() to return -EINVAL on unrecognised command
eeb [Tue, 11 Nov 2003 16:42:34 +0000 (16:42 +0000)]
*   fixed kportal_router_cmd() to return -EINVAL on unrecognised command

21 years agoupdate netconsole-2.4.20-rh.patch
wangdi [Tue, 11 Nov 2003 10:31:39 +0000 (10:31 +0000)]
update netconsole-2.4.20-rh.patch

21 years agoliblustre:
ericm [Tue, 11 Nov 2003 05:18:13 +0000 (05:18 +0000)]
liblustre:
 - support llog facilities
 - remove unused wait_for_completion/completion pairs

21 years ago- we don't have an arch/x86_64 for kmem_cache_validate(), but the code
zab [Tue, 11 Nov 2003 01:09:42 +0000 (01:09 +0000)]
- we don't have an arch/x86_64 for kmem_cache_validate(), but the code
  doesn't use it anymore, so we just won't include it in the series.

21 years ago- async deletion for suse kernels
alex [Tue, 11 Nov 2003 00:35:13 +0000 (00:35 +0000)]
- async deletion for suse kernels

21 years ago- suse series update
alex [Tue, 11 Nov 2003 00:19:18 +0000 (00:19 +0000)]
- suse series update

21 years ago- remove ptlrpc_lib.c. The remaining functions there were moved to
rread [Mon, 10 Nov 2003 18:44:06 +0000 (18:44 +0000)]
- remove ptlrpc_lib.c.  The remaining functions there were moved to
ldlm_lib.c when b_zcfg landed.

21 years agob_2075
youfeng [Mon, 10 Nov 2003 09:36:30 +0000 (09:36 +0000)]
b_2075
make portals_debug_dumpstack() also can be used for vanilla-2.4.20 serie on
i386 host

21 years agob=2005
wangchao [Mon, 10 Nov 2003 02:39:41 +0000 (02:39 +0000)]
b=2005
r=Peter

Add a check for current->fsuid == 0, to protect portals device from non-root usage.

21 years ago- the ancient fstat/fstat64 wart does not warrant a #warning
zab [Sat, 8 Nov 2003 01:21:37 +0000 (01:21 +0000)]
- the ancient fstat/fstat64 wart does not warrant a #warning

21 years agoUp MAXARGS to 512, so that we can have > 95 OSTs visible to lov_setup.
shaver [Fri, 7 Nov 2003 21:55:22 +0000 (21:55 +0000)]
Up MAXARGS to 512, so that we can have > 95 OSTs visible to lov_setup.

21 years ago- set default timeout at 10, so I can run this without timing out on UML
rread [Fri, 7 Nov 2003 21:15:05 +0000 (21:15 +0000)]
- set default timeout at 10, so I can run this without timing out on UML

21 years ago- remove use of "lctl device_list" from lconf, now using
rread [Fri, 7 Nov 2003 18:52:24 +0000 (18:52 +0000)]
- remove use of "lctl device_list" from lconf, now using
    /proc/fs/lustre/devices

21 years agob=2225 fix obd_self_export issues
rread [Fri, 7 Nov 2003 18:52:15 +0000 (18:52 +0000)]
b=2225 fix obd_self_export issues

21 years ago- convert the IOC_*_ROUTE constants to to NAL_CMDs so they are now
rread [Fri, 7 Nov 2003 18:09:56 +0000 (18:09 +0000)]
- convert the IOC_*_ROUTE constants to to NAL_CMDs so they are now
just integers and 64bit safe.

21 years agoBump kernel version to 30 because of portals_debug_dumpstack() kernel patch.
adilger [Fri, 7 Nov 2003 17:25:52 +0000 (17:25 +0000)]
Bump kernel version to 30 because of portals_debug_dumpstack() kernel patch.
Put workaround in place until l30 kernels are available so we can work.

21 years agoAdd define of LCONF LMC LCTL in replay-dual.sh replay-single.sh replay-ost-single...
sice [Fri, 7 Nov 2003 08:22:15 +0000 (08:22 +0000)]
Add define of LCONF LMC LCTL in replay-dual.sh replay-single.sh replay-ost-single.sh test-framework.sh

21 years agofix the malform error in kernel_patches/patches/kernel_text_address-2.4.20-vanilla...
youfeng [Fri, 7 Nov 2003 08:17:27 +0000 (08:17 +0000)]
fix the malform error in kernel_patches/patches/kernel_text_address-2.4.20-vanilla.patch of my previous patch for b_2075

21 years agoromve the reduntant/useless modifications in kernel_patches/patches/kernel_text_addre...
youfeng [Fri, 7 Nov 2003 07:06:11 +0000 (07:06 +0000)]
romve the reduntant/useless modifications in kernel_patches/patches/kernel_text_address-2.4.20-rh.patch on my bugfix 2075

21 years agob_2075
youfeng [Fri, 7 Nov 2003 03:56:29 +0000 (03:56 +0000)]
b_2075
print out the call stack

21 years ago* Added .cvsignore
eeb [Wed, 5 Nov 2003 09:23:11 +0000 (09:23 +0000)]
*   Added .cvsignore

*   Fixed typo in Makefile.am

21 years ago- change NETWORKTYPE to NETTYPE, and make sure it can be overridden.
rread [Wed, 5 Nov 2003 08:35:47 +0000 (08:35 +0000)]
- change NETWORKTYPE to NETTYPE, and make sure it can be overridden.

21 years agob=2004
rread [Wed, 5 Nov 2003 06:18:18 +0000 (06:18 +0000)]
b=2004

There is enough zeroconf infrastructure in place to enable replay-dual
to be used. It fails the first test, though.

21 years agor=2152
rread [Wed, 5 Nov 2003 05:57:46 +0000 (05:57 +0000)]
r=2152

replay-single.sh now supports multiple nodes. The script must be run
on the client, but the mds, mdsfailover, and ost can all be different
nodes. If mdsfailover_HOST is set, then the MDS service  will be
failed between the two mds nodes.

uml1# PDSH="pdsh -w" mds_HOST=uml2 ost_HOST=uml3 mdsfailover_HOST=uml4
  ./replay-single.sh

It still runs on a single node as before.

21 years agoliblustre: minor fix to shared lib creation.
ericm [Wed, 5 Nov 2003 02:28:44 +0000 (02:28 +0000)]
liblustre: minor fix to shared lib creation.

21 years agob=1028
phil [Wed, 5 Nov 2003 00:49:25 +0000 (00:49 +0000)]
b=1028
The totally non-controversial parts of my outstanding mds epoch code.

- add I/O epoch fields to the mds_body, mds_obd, and ll_inode_info
- store the epoch in the LLI after a successful open
- add a data structure, which will go into an MDS inode's i_filterdata

21 years ago* Addid portals/knals/ibnal
eeb [Tue, 4 Nov 2003 19:50:20 +0000 (19:50 +0000)]
*   Addid portals/knals/ibnal

21 years ago* Removed dead code
eeb [Tue, 4 Nov 2003 19:49:40 +0000 (19:49 +0000)]
*   Removed dead code

21 years agoliblustre: improve the makefile. and generation of shared libs.
ericm [Tue, 4 Nov 2003 04:36:16 +0000 (04:36 +0000)]
liblustre: improve the makefile. and generation of shared libs.

21 years agoadd two -ETXTBSY tests, and a pre-creation OST vs. MDS test
phil [Tue, 4 Nov 2003 00:08:47 +0000 (00:08 +0000)]
add two -ETXTBSY tests, and a pre-creation OST vs. MDS test

I hope sanity.sh is not expected to work on multi-node setups?

21 years ago- vanilla-2.4.22 series
alex [Tue, 4 Nov 2003 00:05:02 +0000 (00:05 +0000)]
- vanilla-2.4.22 series

21 years agoDon't do setattr after transaction handle has been committed.
adilger [Mon, 3 Nov 2003 23:41:31 +0000 (23:41 +0000)]
Don't do setattr after transaction handle has been committed.
Combine size and timestamp setattrs, and update size under i_sem.
Use client timestamps instead of server timestamps for files.

21 years agoPrint a more useful error if we have problems cleaning up the portals slabs.
adilger [Mon, 3 Nov 2003 22:57:30 +0000 (22:57 +0000)]
Print a more useful error if we have problems cleaning up the portals slabs.
I don't know why this would fail (we check refcounts above), but it did.

21 years ago* Stale socknal connections culled when a rebooted node reconnects
eeb [Mon, 3 Nov 2003 16:48:18 +0000 (16:48 +0000)]
*    Stale socknal connections culled when a rebooted node reconnects
     (Bug 2062)

*    Removed all userspace HELLO handshake (== nid exchange) code.  It's
     all done by 1 procedure in the kernel now.

*    Portals protocol version up: 0.3

21 years agoliblustre:
ericm [Sat, 1 Nov 2003 09:23:51 +0000 (09:23 +0000)]
liblustre:
 - support rename() & related sanity test
 - fixes in getattr, according to libsysio changes

21 years agoliblustre:
ericm [Sat, 1 Nov 2003 08:17:23 +0000 (08:17 +0000)]
liblustre:
 - support hard link()
 - fix memory leak in getattr()

21 years agoliblustre: follow the libsysio changes
ericm [Sat, 1 Nov 2003 03:12:20 +0000 (03:12 +0000)]
liblustre: follow the libsysio changes

21 years agoland zcfg on devel
rread [Fri, 31 Oct 2003 19:39:23 +0000 (19:39 +0000)]
land zcfg on devel

- includes changes from b_llogging

- the MDS -> LOV connection is created using the MDS config log, so the
  log must exist. Lconf will create the log automatically when
  --reformat is used. To create the config logs on an existing
  filesystem, run lconf on the MDS with --write_conf. This only needs to
  be done on the MDS.

- LOV does not connect to the MDS during setup.  Instead, the MDS and
  MDC use obd_get_info("lovdesc") to get the stripe info. The LOVDESC
  and LOVTGTS files on the MDS are no longer used (and not created
  on new filesystems.)

- Zeroconf clients are new support by lconf --zeroconf, and
  replay-single.sh uses this to mount the client.  The exact arguments
  needed for zeroconf will be changing quickly, so don't use
  lconf --zeroconf in other scripts yet.  Instead, once the dust
  clears, lconf will be changed to always do zeroconf mounts, and
  eventually when don't need lconf anymore is when we start changing
  the test scripts.

21 years agoadd crash support and fix some bugs in netconsole
wangdi [Fri, 31 Oct 2003 09:52:15 +0000 (09:52 +0000)]
add crash support and fix some bugs in netconsole

21 years agoadd lov-sanity.sh for extra EA test
sice [Fri, 31 Oct 2003 06:55:51 +0000 (06:55 +0000)]
add lov-sanity.sh for extra EA test

21 years agoPrint out all failures in wiretest instead of just the first one.
adilger [Thu, 30 Oct 2003 22:57:00 +0000 (22:57 +0000)]
Print out all failures in wiretest instead of just the first one.
Don't automatically build wiretest.c (it should only be rebuilt rarely).
b=1343,1821

21 years agoAdded a new manpage for lwizard.
radhika [Thu, 30 Oct 2003 18:43:27 +0000 (18:43 +0000)]
Added a new manpage for lwizard.

21 years agoThis file contains few generic functions and used by lhealth and lconf.
sravi [Thu, 30 Oct 2003 15:55:02 +0000 (15:55 +0000)]
This file contains few generic functions and used by lhealth and lconf.
find_prog moved to this file, so lhealth and lconf are using the same.

21 years agoAdd find_irq_function, so we do not need add any ethernet driver patch
wangdi [Thu, 30 Oct 2003 14:09:24 +0000 (14:09 +0000)]
Add find_irq_function, so we do not need add any ethernet driver patch

21 years agoadd netconsole patch for rh-2.4.20
wangdi [Thu, 30 Oct 2003 09:35:34 +0000 (09:35 +0000)]
add netconsole patch for rh-2.4.20

21 years agoadd netconsole 2.4.20-rh patch
wangdi [Thu, 30 Oct 2003 09:30:05 +0000 (09:30 +0000)]
add netconsole 2.4.20-rh patch

21 years ago- put it more consistent debug message that track pages through llite.
zab [Thu, 30 Oct 2003 01:44:55 +0000 (01:44 +0000)]
- put it more consistent debug message that track pages through llite.
  Hopefully it will shed some light on bug 2192.

21 years ago- ext3-inode-reuse patch agaist all the supported series (bug 2185)
alex [Wed, 29 Oct 2003 23:30:22 +0000 (23:30 +0000)]
- ext3-inode-reuse patch agaist all the supported series (bug 2185)

NOTE: still under testing. therefore they aren't included in series

21 years ago- Suse series introduced
alex [Wed, 29 Oct 2003 23:21:56 +0000 (23:21 +0000)]
- Suse series introduced

NOTE: it builds, but isn't tested yet

21 years agoAdded a manpage for lfs and some formatting changes to the lctl manpage.
radhika [Wed, 29 Oct 2003 22:47:06 +0000 (22:47 +0000)]
Added a manpage for lfs and some formatting changes to the lctl manpage.

21 years agoCommit wiretest.c as built on i386 for use on other platforms, instead of
adilger [Wed, 29 Oct 2003 22:25:56 +0000 (22:25 +0000)]
Commit wiretest.c as built on i386 for use on other platforms, instead of
(duh) building it on each platform.
b=1343

21 years ago* Cleaned up ptl_hdr_t definition; removed unused fields and moved the
eeb [Wed, 29 Oct 2003 20:15:37 +0000 (20:15 +0000)]
*   Cleaned up ptl_hdr_t definition; removed unused fields and moved the
    common length field into the common part of the header.

*   Socknal connection handshake no longer optional (it was --nid_exchange
    in lconf, the 'x' option to lclt::[auto_]connect, the -x option in
    acceptor etc).

*   Added 'uname -a' output and gcc version info in a comment to the output
    of portals wirecheck.

*   Bumped up the portals protocol version number.

*   Added "typed" socknal connections to establish either a single
    connection...

 'A'    bidirectional control + bulk

    ...or 3 connections...

         'C'    bidirection control
         'I'    bulk in
         'O'    bulk out

    Individual connections can be created using the relevent flag in
    lctl::connect.

    Autoconnect defaults to creating three CIO connections, but can create
    a single A connection if /proc/sys/socknal/typed is zero when the
    connection is established (CAVEAT EMPTOR this also causes ALL existing
    connections to be driven bi-directionally).

    Note that connections types are determined by the initiator.  The
    connection handshake tells the acceptor what type of connection she has
    just accepted, and she tells her socknal.

*   /proc/sys/socknal/eager_ack is now a bitmap of socket types to do eager
    ack on....

          1    enable on 'A' connections
          2    enable on 'C' connections
          4    enable on 'I' connections
          8    enable on 'O' connections (if any messages are ever received)

    Default: 0

    I doubt this facility will be of much use...

*   /proc/sys/socknal/min_bulk controls the size at which messages are
    considered 'bulk' and preferrential sent on 'C' rather than 'O'
    connections.

*   Changed error handling on connections so that when an error is detected
    on a connection...

       (a) ALL connections to the same NID at the same IP are removed.

       (b) The autoconnect route that generated the connection becomes the
           least favourite so that the next autoconnect attempt will use a
           different IP if redundant autoconnect routes have been
           specified.

21 years agoManpage for lctl.
radhika [Wed, 29 Oct 2003 20:11:11 +0000 (20:11 +0000)]
Manpage for lctl.

21 years agoMan pages for lmc and lconf.
radhika [Wed, 29 Oct 2003 17:51:06 +0000 (17:51 +0000)]
Man pages for lmc and lconf.

21 years agob=2200: perform delorphan recovery when an OST is reintegrated.
shaver [Wed, 29 Oct 2003 17:35:52 +0000 (17:35 +0000)]
b=2200: perform delorphan recovery when an OST is reintegrated.
        - use generic per-obd notification system
- relay LOV's notifications to MDS
- support running delorphan on only one OST, if an OST
  UUID is specified
- suppress CERRORs and refailing for -EIO in interpret_create,
  because we handle it more gracefully now

r=phik.

21 years agoliblustre fixes during the last merge
ericm [Wed, 29 Oct 2003 06:22:20 +0000 (06:22 +0000)]
liblustre fixes during the last merge

21 years agob_2036
youfeng [Wed, 29 Oct 2003 03:47:13 +0000 (03:47 +0000)]
b_2036
formats nids for readability, from "nid" to "nid (format)"
nid    : using LPX64 everywhere
format : "%u:%d.%d.%d.%d" for IP NIDS (leading part iscluster ID)
         "%u:%u" for non-IP NIDs (leading part is cluster ID)

21 years agoAdd wiretest program to do struct size/alignment testing in userspace.
adilger [Tue, 28 Oct 2003 21:16:37 +0000 (21:16 +0000)]
Add wiretest program to do struct size/alignment testing in userspace.
b=1343

21 years agoUpdate wirecheck.c with new structs, fields. Put some of the constants with
adilger [Tue, 28 Oct 2003 21:03:21 +0000 (21:03 +0000)]
Update wirecheck.c with new structs, fields.  Put some of the constants with
the structs that use those constants.
b=1343

21 years agosome changes related to llogging, and now it can compile with enable-orphans on
tianying [Tue, 28 Oct 2003 11:38:06 +0000 (11:38 +0000)]
some changes related to llogging, and now it can compile with enable-orphans on

21 years agoInclude lustre_user.h for user-space interface, and put that in the install
adilger [Tue, 28 Oct 2003 03:37:18 +0000 (03:37 +0000)]
Include lustre_user.h for user-space interface, and put that in the install
package instead of the (deprecated for apps) lustre_lite.h and lustre_idl.h.

21 years agoHave local-large-inode.sh just call local.sh with options instead of
adilger [Mon, 27 Oct 2003 20:13:25 +0000 (20:13 +0000)]
Have local-large-inode.sh just call local.sh with options instead of
duplicating code.  The former didn't work for acc-sm previously anyways,
as it didn't build local-large-inode.xml but rather only local.xml.

21 years agoAdd new LOV EA format which gives us fields for features needed in the post
adilger [Mon, 27 Oct 2003 19:34:19 +0000 (19:34 +0000)]
Add new LOV EA format which gives us fields for features needed in the post
1.0 stage (e.g. RAID, OST migration/replacement) as well as more efficient
storage than the current layout when there are few stripes on lots of OSTs.
As well, we will now be able to back up/restore/manipulate EA data via the
"getfattr" and "setfattr" tools (attr RPM) that are part of EA tools from
the "trusted.lov" attribute (only accessible to root/priviledged users).

Also contains code to transparently migrate from old EA format to new one
the first time the EA is accessed.  Note that there is no going back to old
code on the same filesystem once the EAs have been updated.  EA conversion
is (hopefully) "fail safe" in that an error during conversion just means
that the on-disk EA is kept in old format (clients understand both because of
common code).

Some of the "compat" code can be removed once we hit 1.0 and users have
migrated, but the infrastructure itself should probably remain for the
next time we need to change the EA format.

This also adds "lustre_user.h" which was in b_llp_hp and has existed in
b_llnl for some time, but alas the user EA format had to change in order
to allow specification of striping format.  A recompile should fix it.

Some minor warts remain:
- mds_objids_from_lmm() knows too much about lmm format.  It had an existing
  bug (looping to ld_tgt_count vs lmm_ost_count) and caused another bug for
  me because it does its own unpacking of the lmm) and it would be nice to
  just get rid of it and use obd_unpackmd(lmm) to get an lsm instead and use
  that instead of "ids"
- need to fix up endianness from ll_lov_setstripe() because this is now not
  done on the same host any more (object creation is done on MDS now)
- liblustreapi() should not be doing the printing of structs, but should
  instead just return structs to user-space and/or using a callback function
  supplied by the caller (e.g. for lfind iterating over filesystem).  The
  lov_dump_user_lmm* functions should probably be exported from the library
  and be used as default print functions as a starting point.
b=2097
r=braam

21 years agoAdd mdc_llog_setup and mdc_llog_cleanup methods.
tianying [Mon, 27 Oct 2003 14:20:59 +0000 (14:20 +0000)]
Add mdc_llog_setup and mdc_llog_cleanup methods.

21 years ago- all obd operations should return int
braam [Sat, 25 Oct 2003 19:19:41 +0000 (19:19 +0000)]
- all obd operations should return int
- in the process capture a double free in a prep_async_page error case
  in rw.c
- remove unused logging code
- add a mini fix from Tian to not fail in llog_obd_setup when
  re-entering it.

21 years ago- local-large-inode conf added in acceptance-small.sh
alex [Sat, 25 Oct 2003 18:12:27 +0000 (18:12 +0000)]
- local-large-inode conf added in acceptance-small.sh
- lmc & lconf changed a bit to support mkfsoptions attribute/option

21 years agoFix memory not being freed correctly and incorrect checks aroun memory
green [Sat, 25 Oct 2003 08:17:27 +0000 (08:17 +0000)]
Fix memory not being freed correctly and incorrect checks aroun memory
allocations.

21 years agoAdded a manpage for lfs, some other minor updates to a few chapters.
radhika [Fri, 24 Oct 2003 22:46:34 +0000 (22:46 +0000)]
Added a manpage for lfs, some other minor updates to a few chapters.

21 years agob=1933
phil [Fri, 24 Oct 2003 21:50:03 +0000 (21:50 +0000)]
b=1933
The socknal transmit thread calls into TCP, which will block
allocating memory, and deadlock us if kswapd is waiting on Lustre to
write to OSTs.  Set PF_MEMALLOC on that task.