glusterfs.git/libglusterfs, branch v3.3.0qa39

rebalance: handshake_event_notify to make fsd talk to glusterd

2012-04-25T11:05:35+00:00

Event_notify can be used by others to communicate with glusterd.
A cbk event is also added for future use.

req has a op, and dict.
rsp has op_ret, op_errno, and dict.

With this, rebalance process can update the status before
exiting.

Signed-off-by: shishir gowda 
Change-Id: If5c0ec00514eb3a109a790b2ea273317611e4562
BUG: 807126
Reviewed-on: http://review.gluster.com/3013
Tested-by: Gluster Build System 
Reviewed-by: Amar Tumballi 
Reviewed-by: Vijay Bellur

glusterd: command check failure fixes

2012-04-24T20:32:00+00:00

due to the patch 29f2de478cc6a475e6ae760d9cbe7ac847e9d79c (coverity
issues fix), geo-replications assumption of anything other than -2
is valid error (ie, expected error), was broken, which resulted in
staging failure of all geo-replication commands. now fixed!!!

Change-Id: I26c3574e966dd8ba89ef423a3ad17aa863a62c13
Signed-off-by: Amar Tumballi 
BUG: 815613
Reviewed-on: http://review.gluster.com/3219
Tested-by: Gluster Build System 
Reviewed-by: Anand Avati

statedump: fix deadlock during state dump of fds

2012-04-24T16:44:36+00:00

Existing state dump of FD context tries to be extra safe by trying
to call the fd dump callback outside the inode lock. It acheives
this by taking an fd ref and unreffing it later. This exercise can
be harmful at times when the fd unref performed by state dump ends
up being the last unref and triggers fd_destroy. fd_destroy in turn
triggers inode_unref which blocks on inode table lock, while the
inode table lock was already held by the thread before it even
attempted fd ctx dump.

The fix takes away the dangerous ref/unref of the fd during state
dump and instead calls fd_ctx_dump() whiel the inode lock is held.
This is not a problem as long as the dump functions do not call any
inode function which tries to take an inode lock (none of the
existing fd dump ops do)

Change-Id: Ia4b27ba5a321285d3e64896a679a00363bb1e822
BUG: 815242
Signed-off-by: Anand Avati 
Reviewed-on: http://review.gluster.com/3210
Tested-by: Gluster Build System 
Reviewed-by: Amar Tumballi 
Reviewed-by: Raghavendra Bhat 
Reviewed-by: Vijay Bellur

libglusterfs/inode.c: add inode_resolve

2012-04-24T07:06:39+00:00

inode_resolve takes an itable and path as arguments and returns either
NULL or the inode corresponding to the path

Change-Id: Id42d62cce6b04dbfec7d606120b09a1e54ab484e
BUG: 802905
Signed-off-by: Raghavendra G 
Reviewed-on: http://review.gluster.com/3021
Tested-by: Gluster Build System 
Reviewed-by: Raghavendra Bhat 
Reviewed-by: Mohammed Junaid 
Reviewed-by: Anand Avati

NetBSD build fixes

2012-04-24T04:29:03+00:00

Signed-off-by: Emmanuel Dreyfus 

Change-Id: I8f9aabeadd2f842521a82e59594115bd80155d68
BUG: 2923
Signed-off-by: Emmanuel Dreyfus 
Reviewed-on: http://review.gluster.com/3053
Tested-by: Gluster Build System 
Reviewed-by: Anand Avati

core: coverity issues fixed

2012-04-23T21:52:57+00:00

this is not a complete set of issues getting fixed. Will
address other issues in another patch.

Change-Id: Ib01c7b11b205078cc4d0b3f11610751e32d14b69
Signed-off-by: Amar Tumballi 
BUG: 789278
Reviewed-on: http://review.gluster.com/3145
Tested-by: Gluster Build System 
Reviewed-by: Jeff Darcy

clean up handling of special client pids

2012-04-23T21:50:55+00:00

- no need to name a lower boundary [drop GF_CLIENT_PID_MIN]
- drop utility function to decide about pid speciality [gf_pid_client_check()];
  when we implement uniform policy for special pids, use just comparison with
  GF_CLIENT_PID_MAX (currently such uniform policy exists in io-threads, where
  special clients get lowest priority)
- io-threads: drop numeric literal, use GF_CLIENT_PID_MAX
- xattr flipping (in fuse):
  - fuse_xattr_alloc_default() made static
  - avoid compiler warnings wrt. partial enum traversal in the switch
    by not casting the original numeric value to gf_client_pid_t (enum) type
  - trim down the switch to those cases with which there is something to do
    [GF_CLIENT_PID_GSYNCD, GF_CLIENT_PID_HADOOP]
  - eliminate the (theoretical) chance of succesfully returning
    from fuse_flip_xattr_ns() with unallocated nkey
    [as was the case with pids GF_CLIENT_PID_{MIN,MAX,DEFRAG}]

Change-Id: Ic2b83b908d8e3be23a7fba1e5895869828a6194b
BUG: 765433
Signed-off-by: Csaba Henk 
Reviewed-on: http://review.gluster.com/3170
Tested-by: Gluster Build System 
Reviewed-by: Jeff Darcy 
Reviewed-by: Anand Avati

Make option code quieter to make logs more readable.

2012-04-23T21:46:46+00:00

A lot of this stuff is extremely uninformative, such as messages about
default values that can be read from the code.  Having them in the logs
makes more important messages harder to find, so they're down to
GF_DEBUG_TRACE now.

Change-Id: I18921b3deac7a01ad94c50670b90342d1130837b
Signed-off-by: Jeff Darcy 
Reviewed-on: http://review.gluster.com/3198
Tested-by: Gluster Build System 
Reviewed-by: Kaleb KEITHLEY 
Reviewed-by: Anand Avati

libglusterfs: Never block syncproc

2012-04-23T20:11:49+00:00

Change-Id: I64cd8a2ef37926173c19a33df0716183530e22bf
BUG: 814074
Signed-off-by: Pranith Kumar K 
Reviewed-on: http://review.gluster.com/3194
Tested-by: Gluster Build System 
Reviewed-by: Jeff Darcy 
Reviewed-by: Anand Avati

cli: Strip whitespace from "volume set" option values

2012-04-20T16:32:17+00:00

Strips the whitespace from options values before sending to glusterd.
This prevents options containing whitespace to be written to the volfiles which
would cause the volfile parser to fail.

Change-Id: I46faee7b0853141fa102d06bb067c7ab499a2f6e
BUG: 813937
Signed-off-by: Kaushal M 
Reviewed-on: http://review.gluster.com/3192
Tested-by: Gluster Build System 
Reviewed-by: Amar Tumballi 
Reviewed-by: Vijay Bellur