glusterfs.git/tests, branch v3.8.7

cluster/afr: Fix per-txn optimistic changelog initialisation

2016-12-13T09:03:15+00:00

        Backport of: http://review.gluster.org/16075

Incorrect initialisation of local->optimistic_change_log was leading
to skipped pre-op and post-op even when a brick didn't participate in
the txn because it was down.
The result - missing granular name index resulting in some entries
never getting healed.

FIX:
Initialise local->optimistic_change_log just before pre-op.

Also fixed granular entry heal to create the granular name index in
pre-op as opposed to post-op. This is to prevent loss of granular
information when during an entry txn, the good (src) brick goes
offline before the post-op is done. This would cause self-heal to
do conservative merge (since dirty xattr is the only information
available), which when granular-entry-heal is enabled, expects
granular indices, the lack of which can lead to loss of data in
the worst case.

Change-Id: Ibc0fbfb3fa21c578e28868d9e30b274e33c12064
BUG: 1403646
Signed-off-by: Krutika Dhananjay 
Reviewed-on: http://review.gluster.org/16105
Reviewed-by: Pranith Kumar Karampuri 
Smoke: Gluster Build System 
NetBSD-regression: NetBSD Build System 
CentOS-regression: Gluster Build System

syncop: fix conditional wait bug in parallel dir scan

2016-12-12T07:16:47+00:00

Problem:
The issue as seen by the user is detailed in the BZ but what is
happening is if the no. of items in the wait queue == max-qlen,
syncop_mt_dir_scan() does a pthread_cond_wait until the launched
synctask workers dequeue the queue. But if for some reason the worker
fails, the queue is never emptied due to which further invocations of
syncop_mt_dir_scan() are blocked forever.

Fix: Made some changes to _dir_scan_job_fn

- If a worker encounters error while processing an entry, notify the
  readdir loop in syncop_mt_dir_scan() of the error but continue to process
  other entries in the queue, decrementing the qlen as and when we dequeue
  elements, and ending only when the queue is empty.

- If the readdir loop in syncop_mt_dir_scan() gets an error form the
  worker, stop the readdir+queueing of further entries.

> Reviewed-on: http://review.gluster.org/16073
> Smoke: Gluster Build System 
> NetBSD-regression: NetBSD Build System 
> CentOS-regression: Gluster Build System 
> Reviewed-by: Pranith Kumar Karampuri 
(cherry picked from commit 2d012c4558046afd6adb3992ff88f937c5f835e4)

Change-Id: I39ce073e01a68c7ff18a0e9227389245a6f75b88
BUG: 1403192
Signed-off-by: Ravishankar N 
Reviewed-on: http://review.gluster.org/16096
Smoke: Gluster Build System 
NetBSD-regression: NetBSD Build System 
CentOS-regression: Gluster Build System 
Reviewed-by: Pranith Kumar Karampuri

uss: snapd should enable SSL if SSL is enabled on volume

2016-12-11T10:09:47+00:00

During snapd graph generation we should check if SSL is
enabled on main volume or not. This is because clients
will communicate with snapd as if it is communicating to
a brick.


> Reviewed-on: http://review.gluster.org/15979
> Smoke: Gluster Build System 
> NetBSD-regression: NetBSD Build System 
> CentOS-regression: Gluster Build System 
> Reviewed-by: Kaushal M 
(cherry picked from commit 182f0d12040dab5081ca645a3f370f65cd68b528)

Change-Id: I0d7fe86c567b297a8528a48faf06161d4c3cb415
Signed-off-by: Rajesh Joseph 
BUG: 1400459
Reviewed-on: http://review.gluster.org/15986
Smoke: Gluster Build System 
CentOS-regression: Gluster Build System 
NetBSD-regression: NetBSD Build System 
Reviewed-by: Atin Mukherjee

cluster/afr: CLI for granular entry heal enablement/disablement

2016-11-29T10:52:25+00:00

        Backport of: http://review.gluster.org/15747
When there are already existing non-granular indices created that are
yet to be healed, if granular-entry-heal option is toggled from 'off' to
'on', AFR self-heal whenever it kicks in, will try to look for granular
indices in 'entry-changes'. Because of the absence of name indices,
granular entry healing logic will fail to heal these directories, and
worse yet unset pending extended attributes with the assumption that
are no entries that need heal.

To get around this, a new CLI is introduced which will invoke glfsheal
program to figure whether at the time an attempt is made to enable
granular entry heal, there are pending heals on the volume OR there
are one or more bricks that are down. If either of them is true, the
command will be failed with the appropriate error.

New CLI: gluster volume heal  granular-entry-heal {enable,disable}

Change-Id: I342e0390f847fcb015a50ef58aedfcbcb58f4ed3
BUG: 1398501
Signed-off-by: Krutika Dhananjay 
Reviewed-on: http://review.gluster.org/15942
NetBSD-regression: NetBSD Build System 
Smoke: Gluster Build System 
CentOS-regression: Gluster Build System 
Reviewed-by: Pranith Kumar Karampuri

libglusterfs: Fix a read hang

2016-11-29T04:02:19+00:00

Issue:
=====
In certain cases, there was no unwind of read
from read-ahead xlator, thus resulting in hang.

RCA:
====
In certain cases, ioc_readv() issues STACK_WIND_TAIL() instead
of STACK_WIND(). One such case is when inode_ctx for that file
is not present (can happen if readdirp was called, and populates
md-cache and serves all the lookups from cache).

Consider the following graph:
...
io-cache (parent)
   |
readdir-ahead
   |
read-ahead
...

Below is the code snippet of ioc_readv calling STACK_WIND_TAIL:
ioc_readv()
{
...
 if (!inode_ctx)
   STACK_WIND_TAIL (frame, FIRST_CHILD (frame->this),
                    FIRST_CHILD (frame->this)->fops->readv, fd,
                    size, offset, flags, xdata);
   /* Ideally, this stack_wind should wind to readdir-ahead:readv()
      but it winds to read-ahead:readv(). See below for
      explaination.
    */
...
}

STACK_WIND_TAIL (frame, obj, fn, ...)
{
  frame->this = obj;
  /* for the above mentioned graph, frame->this will be readdir-ahead
   * frame->this = FIRST_CHILD (frame->this) i.e. readdir-ahead, which
   * is as expected
   */
  ...
  THIS = obj;
  /* THIS will be read-ahead instead of readdir-ahead!, as obj expands
   * to "FIRST_CHILD (frame->this)" and frame->this was pointing
   * to readdir-ahead in the previous statement.
   */
  ...
  fn (frame, obj, params);
  /* fn will call read-ahead:readv() instead of readdir-ahead:readv()!
   * as fn expands to "FIRST_CHILD (frame->this)->fops->readv" and
   * frame->this was pointing ro readdir-ahead in the first statement
   */
  ...
}

Thus, the readdir-ahead's readv() implementation will be skipped, and
ra_readv() will be called with frame->this = "readdir-ahead" and
this = "read-ahead". This can lead to corruption / hang / other problems.
But in this perticular case, when 'frame->this' and 'this' passed
to ra_readv() doesn't match, it causes ra_readv() to call ra_readv()
again!. Thus the logic of read-ahead readv() falls apart and leads to
hang.

Solution:
=========
Modify STACK_WIND_TAIL() as:
STACK_WIND_TAIL (frame, obj, fn, ...)
{
  next_xl = obj /* resolve obj as the variables passed in obj macro
                   can be overwritten in the further instrucions */
  next_xl_fn = fn /* resolve fn and store in a tmp variable, before
                     modifying any variables */
  frame->this = next_xl;
  ...
  THIS = next_xl;
  ...
  next_xl_fn (frame, next_xl, params);
  ...
}

BUG: 1399018
Change-Id: Ie662ac8f18fa16909376f1e59387bc5b886bd0f9
Signed-off-by: Poornima G 
Reviewed-on: http://review.gluster.org/15934
Smoke: Gluster Build System 
NetBSD-regression: NetBSD Build System 
CentOS-regression: Gluster Build System 
Reviewed-by: Raghavendra G

features/index: Delete granular entry indices of already healed directories during crawl

2016-11-26T12:19:08+00:00

        Backport of: http://review.gluster.org/15880

If granular name indices are already in existence for a volume, and
before they are healed, granular entry heal be disabled, a crawl on
indices/xattrop will clear the changelogs on these directories. When
their corresponding entry-changes indices are crawled subsequently,
if it is found that the directories don't need heal anymore, the
granular indices are not cleaned up.
This patch fixes that problem by ensuring that the zero-xattrop
also deletes the stale indices at the level of index translator.

Change-Id: Iae0a560c1c9d37b083cad89f16d3dcf83c4f7dc7
BUG: 1398501
Signed-off-by: Krutika Dhananjay 
Reviewed-on: http://review.gluster.org/15927
Smoke: Gluster Build System 
NetBSD-regression: NetBSD Build System 
CentOS-regression: Gluster Build System 
Reviewed-by: Pranith Kumar Karampuri

glusterd: clean up old port and allocate new one on every restart

2016-11-23T10:37:16+00:00

    Backport of http://review.gluster.org/#/c/15005/9.

GlusterD as of now was blindly assuming that the brick port which was
already allocated would be available to be reused and that assumption
is absolutely wrong.

Solution : On first attempt, we thought GlusterD should check if the
already allocated brick ports are free, if not allocate new port and
pass it to the daemon. But with that approach there is a possibility
that if PMAP_SIGNOUT is missed out, the stale port will be given back
to the clients where connection will keep on failing. Now given the
port allocation always start from base_port, if everytime a new port
has to be allocated for the daemons, the port range will still be
under control. So this fix tries to clean up old port using
pmap_registry_remove () if any and then goes for pmap_registry_alloc ()

This patch is being ported to 3.8 branch because, the brick process
blindly re-using old port, without registering with the pmap server,
causes snapd daemon to not start properly, even though snapd registers
with the pmap server. With this patch, all the brick processes and
snapd will register with the pmap server to either get the same port,
or a new port, and avoid port collision.

> Reviewed-on: http://review.gluster.org/15005
> Smoke: Gluster Build System 
> NetBSD-regression: NetBSD Build System 
> CentOS-regression: Gluster Build System 
> Reviewed-by: Avra Sengupta 

(cherry picked from commit c3dee6d35326c6495591eb5bbf7f52f64031e2c4)

Change-Id: If54a055d01ab0cbc06589dc1191d8fc52eb2c84f
BUG: 1369766
Signed-off-by: Atin Mukherjee 
Reviewed-on: http://review.gluster.org/15308
Tested-by: Avra Sengupta 
NetBSD-regression: NetBSD Build System 
CentOS-regression: Gluster Build System 
Smoke: Gluster Build System 
Reviewed-by: Niels de Vos

marker: Fix inode value in loc, in setxattr fop

2016-11-22T05:43:38+00:00

Backport of http://review.gluster.org/15826

On recieving a rename fop, marker_rename() stores the,
oldloc and newloc in its 'local' struct, once the rename
is done, the xtime marker(last updated time) is set on
the file, but sending a setxattr fop. When upcall
receives the setxattr fop, the loc->inode is NULL and
it crashes. The loc->inode can be NULL only in one valid
case, i.e. in rename case where the inode of new loc
can be NULL. Hence, marker should have filled the inode
of the new_loc before issuing a setxattr.

> Reviewed-on: http://review.gluster.org/15826
> NetBSD-regression: NetBSD Build System 
> CentOS-regression: Gluster Build System 
> Reviewed-by: Kotresh HR 
> Smoke: Gluster Build System 
> Reviewed-by: Rajesh Joseph 
(cherry picked from commit 46e5466850311ee69e6ae9a11c2bba2aabadd5de)

Change-Id: Id638f678c3daaf4a5c29b970b58929d377ae8977
BUG: 1396418
Signed-off-by: Poornima G 
Reviewed-on: http://review.gluster.org/15878
Reviewed-by: Rajesh Joseph 
Smoke: Gluster Build System 
NetBSD-regression: NetBSD Build System 
CentOS-regression: Gluster Build System

gfapi: async fops should unref in callbacks

2016-11-08T07:13:20+00:00

If fd is unref'd at the end of async call then the unref in cbks would
lead to double unref and possible crash. Removing duplicate unrefs.

Added unref only in failure cases.

A simple test case has been added to test async write case. Need to
extend the same for other async APIs too.

Details:
All glfd based calls in libgfapi, except for glfs_open and glfs_close,
behave in the same way. At the start of the operation, they take a ref
on glfd and fd. At the end of the operation, they unref it. Async calls
are a little different as they unref in the cbk function. A successfull
open call does not unref either the glfd or fd, thereby functioning as a
reference for a OPEN file object. glfs_close makes a syncop_flush call
sandwiched between a fd ref and unref(this can be removed, more on this
below), followed by a call to glfs_mark_glfd_for_deletion which unrefs
glfd and also calls glfs_fd_destroy as a release function thereby doing
a unref on fd too.

Functionally, there is no problem with how everything works when as
described above. However, it is a little non-intuitive that we need to
perform a fd_unref as a consequence of a implicit fd_ref that happens
within glfs_resolve_fd. As we perform a GF_REF_GET(glfd) at the start of
every operation, it would be worthwhile to remove the fd_ref that
glfs_resovle_fd takes and do away with explicit fd_unref()s at the end
of every operation. This is the same reason why we don't need the fd_ref
in glfs_close. This is however not in the scope of this patch.

Change-Id: I86b1d3b2ad846b16ea527d541dc82b5e90b0ba85
BUG: 1392288
Signed-off-by: Raghavendra Talur 
Reviewed-on: http://review.gluster.org/15768
Smoke: Gluster Build System 
NetBSD-regression: NetBSD Build System 
Reviewed-by: Rajesh Joseph 
Reviewed-by: Xavier Hernandez 
Reviewed-by: soumya k 
Reviewed-by: Prasanna Kumar Kalever 
CentOS-regression: Gluster Build System 
Reviewed-by: Shyamsundar Ranganathan 
(cherry picked from commit e65738818dd22462ec00dda021566654d1c702b1)
Reviewed-on: http://review.gluster.org/15779

afr,ec: Heal device files with correct major, minor numbers

2016-10-27T06:23:45+00:00

Thanks a lot to xiaoping.wu@nokia.com from Nokia for the bug and the
fix.

 >BUG: 1384297
 >Change-Id: Ie443237e85d34633b5dd30f85eaa2ac34e45754c
 >Signed-off-by: Pranith Kumar K 
 >Reviewed-on: http://review.gluster.org/15728
 >Smoke: Gluster Build System 
 >NetBSD-regression: NetBSD Build System 
 >Reviewed-by: Xavier Hernandez 
 >CentOS-regression: Gluster Build System 

Change-Id: I7646adc3771ff76cdf9c979b575bbcd0b3bc1b9a
BUG: 1388948
Signed-off-by: Pranith Kumar K 
Reviewed-on: http://review.gluster.org/15735
Smoke: Gluster Build System 
NetBSD-regression: NetBSD Build System 
CentOS-regression: Gluster Build System 
Reviewed-by: Xavier Hernandez