IF YOU WOULD LIKE TO GET AN ACCOUNT, please write an
email to Administrator. User accounts are meant only to access repo
and report issues and/or generate pull requests.
This is a purpose-specific Git hosting for
BaseALT
projects. Thank you for your understanding!
Только зарегистрированные пользователи имеют доступ к сервису!
Для получения аккаунта, обратитесь к администратору.
NLM expects a successful F_UNLCK on an fd to set flock.l_type to F_UNLCK
if it were the last unlock else F_RDLCK.
Change-Id: Ib304ac6102664abbb13d1fda649b3da63f1ee97e
BUG: 800300
Signed-off-by: Krishnan Parthasarathi <kp@gluster.com>
Reviewed-on: http://review.gluster.com/3306
Tested-by: Gluster Build System <jenkins@build.gluster.com>
Reviewed-by: Krishna Srinivas <krishna@gluster.com>
Reviewed-by: Anand Avati <avati@redhat.com>
NFS server is restarted unconditionally when a volume option is
configured through the set interface. This patch prevents restart
of NFS server when operations are performed on translators that
are not part of the NFS graph.
This does not prevent re-start of a NFS server when an option
corresponding to a translator that is part of the NFS graph is
re-configured.
Change-Id: Ic4b8e48e5e7e80438f230521042c267ec3b96a25
Signed-off-by: Vijay Bellur <vijay@gluster.com>
Reviewed-on: http://review.gluster.com/3247
Tested-by: Gluster Build System <jenkins@build.gluster.com>
Reviewed-by: Raghavendra Bhat <raghavendrabhat@gluster.com>
Reviewed-by: Anand Avati <avati@redhat.com>
mkdir and rename both first happen on the hashed subvol.
Rmdir should succeed, only if all non hashed subvol succeed,
and finally removal from hashed_subvol succeeds.
if hashed_subvol is unavailable, then rmdir should be sent
on all subvols
Change-Id: I1bf5d5fc25393d0e2fde13ae403cf867a01f5ec7
BUG: 773187
Signed-off-by: shishir gowda <shishirng@gluster.com>
Reviewed-on: http://review.gluster.com/3202
Tested-by: Gluster Build System <jenkins@build.gluster.com>
Reviewed-by: Amar Tumballi <amarts@redhat.com>
Reviewed-by: Anand Avati <avati@redhat.com>
If the mountpoint path given to glusterfs is not an absolute path, convert it to
an absolute path by concatenating it with the curren working directory.
This prevents cases, where in gluster cannot perform clean unmounts when mount
is done with a relative path.
Change-Id: Ie25add4e1dc59171e522c4244c79a6c148844ab3
BUG: 819466
Signed-off-by: Kaushal M <kaushal@redhat.com>
Reviewed-on: http://review.gluster.com/3302
Tested-by: Gluster Build System <jenkins@build.gluster.com>
Reviewed-by: Amar Tumballi <amarts@redhat.com>
Reviewed-by: Anand Avati <avati@redhat.com>
when parent belongs to active itable.
When parent is root, the parent inode returned will always be
active_subvol->itable->root and hence there can be cases where we
should explicitly resolve the entry in active graph.
Change-Id: I4e82df9a351ff6b5304891abc9932495bf7ea79d
BUG: 804592
Signed-off-by: Raghavendra G <raghavendra@gluster.com>
Reviewed-on: http://review.gluster.com/3007
Tested-by: Gluster Build System <jenkins@build.gluster.com>
Reviewed-by: Anand Avati <avati@redhat.com>
The md-cache translator can cache items for a long period of time and
thus is sensitive to memory usage associated with the items it
references. This implements two changes to help conserve memory:
- quick-read - Migrate the file content data to a local dict and
delete from the xdata dict that passes through the stack.
- md-cache - Create a local dict to selectively store the xattr
data md-cache is interested in. This includes a slight
optimization to not allocate an empty dict in the case
where we have not received any xattr's of interest.
I've tested both changes independently and together by running a
readdirp test against several compiled source trees (~340k files).
The base test results in a 7.7GB RSS on the client. The quick-read
modification cuts RSS down to 4.1GB, which is still large due
to md-cache unintelligently caching a large number of empty
dictionaries. The combined modification cuts RSS down to 462MB
on the same workload.
NOTE: Theoretically the md-cache change should supercede the
quick-read change, but practically I save an extra 150MB or so with
both. I already had the change and consider it an improvement that
quick-read clean up after itself.
BUG: 812876
Change-Id: Id59734d12dd6476b0e32480939e633448adb6884
Signed-off-by: Brian Foster <bfoster@redhat.com>
Reviewed-on: http://review.gluster.com/3268
Tested-by: Gluster Build System <jenkins@build.gluster.com>
Reviewed-by: Raghavendra G <raghavendra@gluster.com>
Reviewed-by: Amar Tumballi <amarts@redhat.com>
Reviewed-by: Jeff Darcy <jdarcy@redhat.com>
Make use of gfid in log messages.
Change-Id: I28548be64d19418f99ee9ccdbd3ac4ae67c6c015
BUG: 814666
Signed-off-by: Kaushal M <kaushal@redhat.com>
Reviewed-on: http://review.gluster.com/3274
Tested-by: Gluster Build System <jenkins@build.gluster.com>
Reviewed-by: Vijay Bellur <vijay@gluster.com>
with this, we can make sure, all wrong patterns in CLI can be
reported with proper message back to user, instead of resulting
in an wrong volume type.
Change-Id: Ib461956dd7d51f7f81c19c27112d44c2adb3a66d
Signed-off-by: Amar Tumballi <amarts@redhat.com>
BUG: 803711
Reviewed-on: http://review.gluster.com/3278
Tested-by: Gluster Build System <jenkins@build.gluster.com>
Reviewed-by: Kaushal M <kaushal@redhat.com>
Reviewed-by: Vijay Bellur <vijay@gluster.com>
Lk-owner of posix-lk and flush should be same, flush can't clear
posix-lks without that lk-owner.
Change-Id: If775abb5741a0beb00c419b54d023fbd429e3cb7
BUG: 810502
Signed-off-by: Pranith Kumar K <pranithk@gluster.com>
Reviewed-on: http://review.gluster.com/3221
Tested-by: Gluster Build System <jenkins@build.gluster.com>
Reviewed-by: Vijay Bellur <vijay@gluster.com>
Hence forth, all node related info, which does not need chksum or
versioning can be stored in this file. This data might not be in
sync across nodes, as it contains only information related to
local node.
Change-Id: I88a48c1b7bb7b4c29625550a9ff349442bcc5d09
BUG: 812234
Signed-off-by: shishir gowda <shishirng@gluster.com>
Reviewed-on: http://review.gluster.com/3141
Tested-by: Gluster Build System <jenkins@build.gluster.com>
Reviewed-by: Vijay Bellur <vijay@gluster.com>
- If self-heal is disabled, the crawl is always spawned but immediately
terminated upon realizing that the self-heal-daemon is disabled. Fixed it.
- If for some reason finding child position fails, it should still spawn timer
and should keep the timer alive until the position is found to be REMOTE.
Change-Id: I20cc6af93a1ff26f2b837bbf063bbe508b5fdc78
BUG: 810144
Signed-off-by: Pranith Kumar K <pranithk@gluster.com>
Reviewed-on: http://review.gluster.com/3100
Tested-by: Gluster Build System <jenkins@build.gluster.com>
Reviewed-by: Vijay Bellur <vijay@gluster.com>
provide an option to failover to standard allocation if iobuf of
required size doesn't exists. this can be achieved by keeping an
arena dedicated for all the out of boundary allocations.
Change-Id: I41a2bd7d353dc7bcb2e1a6e4b41735afe9865975
Signed-off-by: Amar Tumballi <amarts@redhat.com>
BUG: 812784
Reviewed-on: http://review.gluster.com/3136
Tested-by: Gluster Build System <jenkins@build.gluster.com>
Reviewed-by: Anand Avati <avati@redhat.com>
Self heal if any will be taken care of the self heal daemon.
Change-Id: I4f3c270b6fce250cbdf6f986066f64e7c4b2f29f
BUG: 808977
Signed-off-by: shishir gowda <shishirng@gluster.com>
Reviewed-on: http://review.gluster.com/3263
Reviewed-by: Amar Tumballi <amarts@redhat.com>
Tested-by: Gluster Build System <jenkins@build.gluster.com>
earlier it was dependant on server sending the dict for knowing
the command. Now it preserves the dictionary in frame->local,
which is used to understand the command.
Change-Id: Ica675b497e466f77878f67792ae96b71f0704d6b
Signed-off-by: Amar Tumballi <amarts@redhat.com>
BUG: 815745
Reviewed-on: http://review.gluster.com/3235
Tested-by: Gluster Build System <jenkins@build.gluster.com>
Reviewed-by: Shishir Gowda <shishirng@gluster.com>
start volume: --first=yes if the volume is the first to be started
and --first=no otherwise
stop volume: --last=yes if the volume is the last to be stopped
and --last=no otherwise
set volume: -o key=value ... for every key, value supplied in
volume set command
Change-Id: Ia8530771f8b9d7424fd1c736cb4c3622b5cabec2
BUG: 806996
Signed-off-by: Krishnan Parthasarathi <kp@gluster.com>
Reviewed-on: http://review.gluster.com/3260
Tested-by: Gluster Build System <jenkins@build.gluster.com>
Reviewed-by: Jeff Darcy <jdarcy@redhat.com>
Reviewed-by: Vijay Bellur <vijay@gluster.com>
As per RFC5531, the value we were using with AUTH_GLUSTERFS_v2 (6)
is already a standard number for RPCSEC_GSS. Hence having a number
from the free pool for now, till IANA approves a standard number
for GlusterFS RPC itself.
Change-Id: I10ef6f349bccf3cf42fadfa3861c374adae2a808
Signed-off-by: Amar Tumballi <amarts@redhat.com>
BUG: 816148
Reviewed-on: http://review.gluster.com/3230
Tested-by: Gluster Build System <jenkins@build.gluster.com>
Reviewed-by: Niels de Vos <ndevos@redhat.com>
Tested-by: Niels de Vos <ndevos@redhat.com>
Reviewed-by: Anand Avati <avati@redhat.com>
When a synctask is executed in synctask_switchto, if by the time
it goes to check the woken/sleep part a reply already comes and
__wake is called, already running task is going to be put in runq
this generates a false warning "re-running already running task".
If the reply does not come before the woken/sleep check, then the
running task is put in waitq which decrements env->runcount even
when the task is not in runq, this leads to -ve runcount everytime
a task goes from runq->switchto->waitq.
This patch fixes both of them by introducing a new state for the
task called SYNCTASK_SUSPEND just when the task is suspended before
yeild in SYNCOP.
Change-Id: Ib82182cf950f9d85b5656f6243541489a104ca3d
BUG: 816551
Signed-off-by: Pranith Kumar K <pranithk@gluster.com>
Reviewed-on: http://review.gluster.com/3249
Tested-by: Gluster Build System <jenkins@build.gluster.com>
Reviewed-by: Vijay Bellur <vijay@gluster.com>
* in posix we log occassionally if errno is ENOTSUP, added a
suggestion to mount with 'user_xattr' option.
* changed server's *etxattr_cbk to log ENOTSUP in debug level.
* changed client's *etxattr_cbk to log ENOTSUP in debug level.
Change-Id: Icd604050aaa68546011f2c950ecd7883ac6ee820
Signed-off-by: Amar Tumballi <amarts@redhat.com>
BUG: 811957
Reviewed-on: http://review.gluster.com/3140
Tested-by: Gluster Build System <jenkins@build.gluster.com>
Reviewed-by: Jeff Darcy <jdarcy@redhat.com>
Reviewed-by: Vijay Bellur <vijay@gluster.com>
Till now if one of the bricks of the volume is down, then statedump of that
volume (bricks) did not take the statedumps of running bricks.
This patch fixes that issue and cli command will take the statedumps of
running brick processes.
Change-Id: Id9b11905d3d759202e505c2ecfa6ccd5dce6a9ff
BUG: 816884
Signed-off-by: Raghavendra Bhat <raghavendrabhat@gluster.com>
Reviewed-on: http://review.gluster.com/3237
Tested-by: Gluster Build System <jenkins@build.gluster.com>
Reviewed-by: Amar Tumballi <amarts@redhat.com>
Reviewed-by: Vijay Bellur <vijay@gluster.com>
While setting fd_ctx, if we have to scale the fd contexts, then we get the
xl_count of the new graph and reallocate the memory for contexts so that
it is suffecient for contexts of xlators of both new as well as old graph, and
then initialize the new graph xlator contexts. For that we are considering the
old pointer itself, instead of using newly reallocated pointer.
Change-Id: I32c18cd69c03c771b4a63e5186c65ae658e54e92
BUG: 817461
Signed-off-by: Raghavendra Bhat <raghavendrabhat@gluster.com>
Reviewed-on: http://review.gluster.com/3250
Tested-by: Gluster Build System <jenkins@build.gluster.com>
Reviewed-by: Amar Tumballi <amarts@redhat.com>
Reviewed-by: Vijay Bellur <vijay@gluster.com>
we were not checking for the return value of the fgetxattr(key), and
used to continue with the allocation even if size was -1, leading to
wrong memory access.
Change-Id: Ib5cf2e74fee95bc919b12efe89fed5cd25807efd
Signed-off-by: Amar Tumballi <amarts@redhat.com>
BUG: 815346
Reviewed-on: http://review.gluster.com/3236
Tested-by: Gluster Build System <jenkins@build.gluster.com>
Reviewed-by: Jeff Darcy <jdarcy@redhat.com>
Reviewed-by: Anand Avati <avati@redhat.com>
get_new_dict does not take a ref. Ref was taken only when any data
was added to the dict.
But in the out tag: we call explicit unref, which would move the ref
count to -1, if it was a unsuccessful call. unref destroys the dict
only if ref == 0.
Change-Id: Ie08c301237c2042daf90a7ef25569e3b06e3e1e9
BUG: 816870
Signed-off-by: shishir gowda <shishirng@gluster.com>
Reviewed-on: http://review.gluster.com/3240
Tested-by: Gluster Build System <jenkins@build.gluster.com>
Reviewed-by: Amar Tumballi <amarts@redhat.com>
Reviewed-by: Jeff Darcy <jdarcy@redhat.com>
Reviewed-by: Vijay Bellur <vijay@gluster.com>
slots for new keys.
Since while migrating fds to new graph we retain the same fd object,
we might run out of slots for new keys in fd->_ctx after some graph
switches.
Change-Id: I1e3865c76f4703768f9b10b0453558877c2f5448
BUG: 811562
Signed-off-by: Raghavendra G <raghavendra@gluster.com>
Reviewed-on: http://review.gluster.com/3201
Tested-by: Gluster Build System <jenkins@build.gluster.com>
Reviewed-by: Amar Tumballi <amarts@redhat.com>
Reviewed-by: Vijay Bellur <vijay@gluster.com>
