glusterfs.git/xlators/cluster/ec, branch v3.7.3

cluster/ec: Handle race between unlock-timer, new lock

2015-07-23T16:36:46+00:00

Problem:
New lock could come at the time timer is on the way to unlock. This was leading
to crash in timer thread because thread executing new lock can free up the
timer_link->fop and then timer thread will try to access structures already
freed.

Fix:
If the timer event is fired, set lock->release to true and wait for unlock to
complete.

Thanks to Xavi and Bhaskar for helping in confirming that this race is the RC.
Thanks to Kritika for pointing out and explaining how Avati's patch can be used
to fix this bug.

> Change-Id: I45fa5470bbc1f03b5f3d133e26d1e0ab24303378
> BUG: 1243187
> Signed-off-by: Pranith Kumar K 
> Reviewed-on: http://review.gluster.org/11670
> Tested-by: Gluster Build System 
> Reviewed-by: Xavier Hernandez 
> Tested-by: NetBSD Build System 

Change-Id: I9af012e717493684b7cd7d1c63baf2fa401fb542
BUG: 1246121
Signed-off-by: Pranith Kumar K 
Reviewed-on: http://review.gluster.org/11752
Tested-by: NetBSD Build System 
Tested-by: Gluster Build System 
Reviewed-by: Xavier Hernandez

cluster/ec: Propogate correct errno in case of failures

2015-07-22T06:28:19+00:00

- Also remove internal-fop setting in create/mknod etc xattrs.

Rebalance was failing because ec was giving EIO when lock acquiring fails as
the file/dir doesn't exist. Posix_create/mknod are not setting config xattr
because internal-fop key is present in dict and setxattr for this fails leading
to failure in setting rest of xattrs.

 >Change-Id: Ifb429c8db9df7cd51e4f8ce53fdf1e1b975c9993
 >BUG: 1242254
 >Signed-off-by: Pranith Kumar K 
 >Reviewed-on: http://review.gluster.org/11639
 >Reviewed-by: Raghavendra G 
 >Tested-by: Gluster Build System 
 >Reviewed-by: Xavier Hernandez 
 >Tested-by: NetBSD Build System 

BUG: 1243654
Change-Id: Iedb90d6a7d980fb88d6dfa6a6c978a165a4be3fd
Signed-off-by: Pranith Kumar K 
Reviewed-on: http://review.gluster.org/11688
Reviewed-by: Xavier Hernandez 
Tested-by: Gluster Build System

cluster/ec: Prevent data corruptions

2015-07-22T06:27:42+00:00

- On lock reuse preserve 'healing' bits
- Don't set ctx->size outside locks in healing code
- Allow xattrop internal fops also on the fop->mask.

 >Change-Id: I6b76da5d7ebe367d8f3552cbf9fd18e556f2a171
 >BUG: 1232678
 >Signed-off-by: Pranith Kumar K 
 >Reviewed-on: http://review.gluster.org/11640
 >Tested-by: NetBSD Build System 
 >Tested-by: Gluster Build System 
 >Reviewed-by: Xavier Hernandez 

BUG: 1243647
Change-Id: I1b3828e4d4a863b84b2c4e732e7965d1302cea47
Signed-off-by: Pranith Kumar K 
Reviewed-on: http://review.gluster.org/11686
Reviewed-by: Xavier Hernandez 
Tested-by: Gluster Build System

cluster/ec: wind readlink on good subvol(s)

2015-07-22T06:27:12+00:00

 >BUG: 1232172
 >Change-Id: I3a56e487840d86147dd85bf5fbe79b165eae289f
 >Signed-off-by: Pranith Kumar K 
 >Reviewed-on: http://review.gluster.org/11589
 >Tested-by: NetBSD Build System 
 >Reviewed-by: Xavier Hernandez 
 >Tested-by: Gluster Build System 

BUG: 1234679
Change-Id: I08560eee095a3921e9c24f16dc2a242a76018a42
Signed-off-by: Pranith Kumar K 
Reviewed-on: http://review.gluster.org/11687
Tested-by: Gluster Build System 
Reviewed-by: Xavier Hernandez

cluster/ec: Don't read from bricks that are healing

2015-07-22T06:26:47+00:00

 >BUG: 1232678
 >Change-Id: I35503039e4723cf7f33d6797f0ba90dd0aca130b
 >Signed-off-by: Pranith Kumar K 
 >Reviewed-on: http://review.gluster.org/11580
 >Tested-by: NetBSD Build System 
 >Tested-by: Gluster Build System 
 >Reviewed-by: Xavier Hernandez 

BUG: 1243647
Change-Id: I4eb45197a5a8d9652eded37ba1e67d9ea745a583
Signed-off-by: Pranith Kumar K 
Reviewed-on: http://review.gluster.org/11685
Reviewed-by: Xavier Hernandez 
Tested-by: Gluster Build System

cluster/ec: Remove locks in opendir

2015-07-21T21:43:40+00:00

With readdir[p] taking locks to figure out which bricks are
good/bad, no need to take any locks on opendir.

 >BUG: 1232172
 >Change-Id: I4d924aeeaecab23af08c4598548a20d2a44cd849
 >Signed-off-by: Pranith Kumar K 
 >Reviewed-on: http://review.gluster.org/11506
 >Reviewed-by: Xavier Hernandez 
 >Tested-by: NetBSD Build System 
 >Tested-by: Gluster Build System 

BUG: 1234679
Change-Id: Ibada9c196fb44a38400ef027f0a3a142d32797a6
Signed-off-by: Pranith Kumar K 
Reviewed-on: http://review.gluster.org/11684
Tested-by: Gluster Build System 
Reviewed-by: Xavier Hernandez

cluster/ec: Fix use after free bug

2015-07-21T21:43:10+00:00

In ec_lock() there is a chance that ec_resume is called on fop even before
ec_sleep.  This can result in refs == 0 for fop leading to use after free in
this function when it calls ec_sleep so do ec_sleep at start and ec_resume at
end of this function.

 >Change-Id: I879b2667bf71eaa56be1b53b5bdc91b7bb56c650
 >BUG: 1240284
 >Signed-off-by: Pranith Kumar K 
 >Reviewed-on: http://review.gluster.org/11558
 >Reviewed-by: Xavier Hernandez 
 >Tested-by: Gluster Build System 
 >Tested-by: NetBSD Build System 

BUG: 1243648
Change-Id: I57515d1f478b2a41a20d37368c947049d23778f0
Signed-off-by: Pranith Kumar K 
Reviewed-on: http://review.gluster.org/11683
Tested-by: Gluster Build System 
Reviewed-by: Xavier Hernandez

cluster/ec: Don't read from bad subvols

2015-07-21T21:42:47+00:00

 >Change-Id: Ic22813371faca4e8198c9b0b20518e68d275f3c1
 >BUG: 1232678
 >Signed-off-by: Pranith Kumar K 
 >Reviewed-on: http://review.gluster.org/11531
 >Reviewed-by: Xavier Hernandez 
 >Tested-by: NetBSD Build System 
 >Tested-by: Gluster Build System 

BUG: 1243647
Change-Id: Ic8570710a16715322bc0be59367007567f6438cd
Signed-off-by: Pranith Kumar K 
Reviewed-on: http://review.gluster.org/11682
Tested-by: Gluster Build System 
Reviewed-by: Xavier Hernandez

cluster/ec: Remove failed subvols from source/sink computation

2015-07-21T21:40:33+00:00

 >Change-Id: Ib0de34c86ee25de361ec821d4015296c514742e0
 >BUG: 1240210
 >Signed-off-by: Pranith Kumar K 
 >Reviewed-on: http://review.gluster.org/11546
 >Reviewed-by: Xavier Hernandez 
 >Tested-by: NetBSD Build System 
 >Tested-by: Gluster Build System 

BUG: 1243644
Change-Id: I41817ca933bb1eecdb3e895a16753226b608814a
Signed-off-by: Pranith Kumar K 
Reviewed-on: http://review.gluster.org/11681
Tested-by: Gluster Build System 
Reviewed-by: Xavier Hernandez

cluster/ec: Make background healing optional behavior

2015-07-21T21:37:38+00:00

Provide options to control number of active background heal count and qlen.

 >Change-Id: Idc2419219d881f47e7d2e9bbc1dcdd999b372033
 >BUG: 1237381
 >Signed-off-by: Pranith Kumar K 
 >Reviewed-on: http://review.gluster.org/11473
 >Reviewed-by: Xavier Hernandez 
 >Tested-by: Gluster Build System 

BUG: 1238476
Change-Id: I22ba902d9911195656db9e458c01b54cf0afcd7a
Signed-off-by: Pranith Kumar K 
Reviewed-on: http://review.gluster.org/11680
Tested-by: Gluster Build System 
Reviewed-by: Xavier Hernandez