glusterfs.git/xlators/cluster/afr/src, branch v3.7.15

cluster/afr: Prevent split-brain when bricks are brought off and on in cyclic order

2016-08-22T10:05:08+00:00

        Backport of: http://review.gluster.org/15080

When the bricks are brought offline and then online in cyclic
order while writes are in progress on a file, thanks to inode
refresh in write txns, AFR will mostly fail the write attempt
when the only good copy is offline. However, there is still a
remote possibility that the file will run into split-brain if
the brick that has the lone good copy goes offline *after* the
inode refresh but *before* the write txn completes (I call it
in-flight split-brain in the patch for ease of reference),
requiring intervention from admin to resolve the split-brain
before the IO can resume normally on the file. To get around this,
the patch does the following things:
i) retains the dirty xattrs on the file
ii) avoids marking the last of the good copies as bad (or accused)
    in case it is the one to go down during the course of a write.
iii) fails that particular write with the appropriate errno.

This way, we still have one good copy left despite the split-brain situation
which when it is back online, will be chosen as source to do the heal.

Change-Id: I7c13c6ddd5b8fe88b0f2684e8ce5f4a9c3a24a08
BUG: 1367270
Signed-off-by: Krutika Dhananjay 
Reviewed-on: http://review.gluster.org/15222
Smoke: Gluster Build System 
NetBSD-regression: NetBSD Build System 
CentOS-regression: Gluster Build System 
Reviewed-by: Oleksandr Natalenko 
Reviewed-by: Pranith Kumar Karampuri

cluster/afr: copy loc before passing to syncop

2016-08-17T11:44:11+00:00

Problem:
When io-threads is enabled on the client side, io-threads destroys the
call-stub in which the loc is stored as soon as the c-stack unwinds.
Because afr is creating a syncop with the address of loc passed in
setxattr by the time syncop tries to access it, io-threads would have
already freed the call-stub. This will lead to crash.

Fix:
Copy loc to frame->local and use it's address.

> Reviewed-on: http://review.gluster.org/15070
> Smoke: Gluster Build System 
> CentOS-regression: Gluster Build System 
> NetBSD-regression: NetBSD Build System 
> Reviewed-by: Ravishankar N 

BUG: 1367305
Change-Id: I16987e491e24b0b4e3d868a6968e802e47c77f7a
Signed-off-by: Pranith Kumar K 
Signed-off-by: Oleksandr Natalenko 
Reviewed-on: http://review.gluster.org/15168
Reviewed-by: Ravishankar N 
Smoke: Gluster Build System 
CentOS-regression: Gluster Build System 
NetBSD-regression: NetBSD Build System

cluster/afr: Bug fixes in txn codepath

2016-08-17T10:22:53+00:00

        Backport of: http://review.gluster.org/15145

AFR sets transaction.pre_op[] array even before actually doing the
pre-op on-disk. Therefore, AFR must not only consider the pre_op[] array
but also the failed_subvols[] information before setting the pre_op_done[]
flag. This patch fixes that.

Change-Id: I8163256a6de254be43a7a526c6d2f9dc30e0e1df
BUG: 1367270
Signed-off-by: Krutika Dhananjay 
Reviewed-on: http://review.gluster.org/15162
Reviewed-by: Ravishankar N 
NetBSD-regression: NetBSD Build System 
Reviewed-by: Pranith Kumar Karampuri 
Smoke: Gluster Build System 
CentOS-regression: Gluster Build System 
Reviewed-by: Anuradha Talur

afr: some coverity fixes

2016-07-28T13:54:36+00:00

Backport of http://review.gluster.org/#/c/14895/

Thanks to Krutika for a cleaner way to track inode refs in
afr_set_split_brain_choice().

Change-Id: I2d968d05b815ad764b7e3f8aa9ad95a792b3c1df
BUG: 1360549
Signed-off-by: Ravishankar N 
Reviewed-on: http://review.gluster.org/15017
Smoke: Gluster Build System 
NetBSD-regression: NetBSD Build System 
CentOS-regression: Gluster Build System 
Reviewed-by: Krutika Dhananjay 
Reviewed-by: Pranith Kumar Karampuri

afr:Don't wind reads for files in metadata split-brain

2016-06-27T07:13:36+00:00

Backport of http://review.gluster.org/#/c/13389/

Problem: For a read on  a file in metadata split-brain:
1.lookup_done resets event_generation to zero.
2. readv is issued, goes to inode refresh due to mismatching event_gen.
3. After refresh is successful, we update event_generation, data and
metdata readable.
3. We then call afr_read_txn_refresh_done() which in turn calls
afr_inode_get_readable() but doesn't check for EIO. So afr_readv_wind
is called with local->readable (which is populated with data_readable),
thus winding the read to a brick.
4. Also, further parallel reads that come directly go to the wind path
because there is no inode_refresh needed.

Fix:
1.For any afr_read_txn(), readable must be an intersection of data and metadata
readable.
2.Check for EIO in afr_read_txn_refresh_done().

Change-Id: I22dd221fdfaf96d7aced2f474e28ed1337d69f0e
BUG: 1349881
Signed-off-by: Ravishankar N 
(cherry picked from commit 7a1c1e2904701496968ed14b6d7479fb706c3188)
Reviewed-on: http://review.gluster.org/14791
NetBSD-regression: NetBSD Build System 
CentOS-regression: Gluster Build System 
Tested-by: Gluster Build System 
Smoke: Gluster Build System 
Reviewed-by: Pranith Kumar Karampuri

dht/afr/client/posix: Fail mkdir without gfid-req

2016-06-17T06:52:55+00:00

Do not allow directory creations without gfids as
after the directories are created, operations
on them fail anyway. So it is better to fail mkdir.

 >BUG: 1317361
 >Change-Id: I8f8e3b38bbded1960b7215bac0432500f7e78038
 >Signed-off-by: Pranith Kumar K 
 >Reviewed-on: http://review.gluster.org/13690
 >Smoke: Gluster Build System 
 >Reviewed-by: Krutika Dhananjay 
 >CentOS-regression: Gluster Build System 
 >NetBSD-regression: NetBSD Build System 
 >(cherry picked from commit b246b07896fefb261c9fb07f3f29f0d03b81b88d)

Change-Id: Ibf9c84add7265e3e1755a37958e1de38307624b2
BUG: 1332372
Signed-off-by: Pranith Kumar K 
Reviewed-on: http://review.gluster.org/14188
Smoke: Gluster Build System 
NetBSD-regression: NetBSD Build System 
CentOS-regression: Gluster Build System

core, shard: Make shards inherit main file's O_DIRECT flag if present

2016-06-16T16:04:10+00:00

        Backport of: http://review.gluster.org/14191

If the application opens a file with O_DIRECT, the shards'
anon fds would also need to inherit the flag. Towards this,
shard xl would be passing the odirect flag in the @flags parameter
to the WRITEV fop. This will be used in anon fd resolution
and subsequent opening by posix xl.

Change-Id: I3a0593fa46cc25e390a5762a0354b469c2a1532d
BUG: 1342903
Signed-off-by: Krutika Dhananjay 
Reviewed-on: http://review.gluster.org/14663
Smoke: Gluster Build System 
CentOS-regression: Gluster Build System 
NetBSD-regression: NetBSD Build System 
Reviewed-by: Jeff Darcy

afr: Consider ENOSPC and EDQUOT as symmetric errors

2016-06-13T06:42:23+00:00

Backport of http://review.gluster.org/#/c/14604/

Problem:
Since commit 8eaa3506ead4f11b81b146a9e56575c79f3aad7b, in replica 3, if a
brick is down and a create fails on the other 2 brick with EDQUOT, we consider
it an unsymmetric error and hence do not do post-op. So the dirty xattr
remains set on the parent dir, leading to conservative merges during heal when
all bricks are up. i.e. a file deleted on the source might re-appear after heal.

Fix:
Consider ENOSPC and  EDQUOT as symmetric errors since there is no
possibility of partial inode or entry modification operations possible when
quota is enabled. IOW, if quota reports EDQUOT, the no. of bytes written
(or not written) will be the same on all bricks of the replica.
Likewise, the entry operation (create, mkdir...) will either succeed or
not succeed on all bricks.

Change-Id: Iacb1108e9ef4a918e36242fb4a957455133744e9
BUG: 1344561
Signed-off-by: Ravishankar N 
Reviewed-on: http://review.gluster.org/14688
Smoke: Gluster Build System 
CentOS-regression: Gluster Build System 
NetBSD-regression: NetBSD Build System 
Reviewed-by: Pranith Kumar Karampuri

cluster/afr: Unwind xdata_rsp even in case of failures

2016-06-03T13:24:00+00:00

DHT expects GF_PREOP_CHECK_FAILED to be present in xdata_rsp in case of mkdir
failures because of stale layout. But AFR was unwinding null xdata_rsp in case
of failures. This was leading to mkdir failures just after remove-brick. Unwind
the xdata_rsp in case of failures to make sure the response from brick reaches
dht.

 >BUG: 1340623
 >Change-Id: Idd3f7b95730e8ea987b608e892011ff190e181d1
 >Signed-off-by: Pranith Kumar K 
 >Reviewed-on: http://review.gluster.org/14553
 >NetBSD-regression: NetBSD Build System 
 >Reviewed-by: Ravishankar N 
 >Smoke: Gluster Build System 
 >CentOS-regression: Gluster Build System 
 >Reviewed-by: Anuradha Talur 
 >Reviewed-by: Krutika Dhananjay 

BUG: 1340992
Change-Id: I2641d35a851be692aa223dfea5d082245ac6c2bc
Signed-off-by: Pranith Kumar K 
Reviewed-on: http://review.gluster.org/14633
NetBSD-regression: NetBSD Build System 
CentOS-regression: Gluster Build System 
Smoke: Gluster Build System

cluster/afr: Don't let NFS cache stat after writes.

2016-06-03T13:23:47+00:00

Problem:
Afr does post-ops after write but the stat buffer it unwinds is at the
time of write, so if nfs client caches this, it will see different
ctime when it does stat on it after post-op is done. From NFS client's
perspective it thinks the file is changed. Tar which depends on this
to be correct keeps giving 'file changed as we read it' warning.
If Afr instead has to choose to unwind after post-op, eager-lock,
delayed-post-op will have to be disabled which will lead to bad
performance for all write usecases.

Fix:
Don't let client cache stat after write.

 >Change-Id: Ic6062acc6e5cdd97a9c83c56bd529ec83cee8a23
 >BUG: 1302948
 >Signed-off-by: Pranith Kumar K 
 >Signed-off-by: Anuradha Talur 
 >Reviewed-on: http://review.gluster.org/13785
 >Smoke: Gluster Build System 
 >NetBSD-regression: NetBSD Build System 
 >CentOS-regression: Gluster Build System 
 >Reviewed-by: Niels de Vos 

BUG: 1312721
Change-Id: I42a5d524bcf2a2034fe48ee8454812ca26a98c37
Signed-off-by: Pranith Kumar K 
Reviewed-on: http://review.gluster.org/14454
NetBSD-regression: NetBSD Build System 
CentOS-regression: Gluster Build System 
Smoke: Gluster Build System