glusterfs-snapshot.git/xlators/protocol/server/src, branch upstream

gNFS: RFE for NFS connection behavior

2013-11-15T00:07:02+00:00

Implement reconfigure() for NFS xlator so that volume set/reset wont
restart the NFS server process. But few options can not be reconfigured
dynamically e.g. nfs.mem-factor, nfs.port etc which needs NFS to be
restarted.

Change-Id: Ic586fd55b7933c0a3175708d8c41ed0475d74a1c
BUG: 1027409
Signed-off-by: Santosh Kumar Pradhan 
Reviewed-on: http://review.gluster.org/6236
Tested-by: Gluster Build System 
Reviewed-by: Rajesh Joseph 
Reviewed-by: Anand Avati

server/rpc: bricks goes offline and comes back online, with lots of "No such file or directory" log messages

2013-11-13T07:38:03+00:00

Problem:
Messages were getting logged very frequently at log level INFO.

[2013-03-01 11:34:28.029222] I [server3_1-fops.c:1541:server_open_cbk] vol-server: 993888: OPEN (null) (--) ==> -1 (No such file or directory)
[2013-03-01 11:34:28.031579] I [server3_1-fops.c:252:server_inodelk_cbk] vol-server: 993896: INODELK (null) (--) ==> -1 (No such file or directory)
[2013-03-01 11:34:28.034041] I [server3_1-fops.c:252:server_inodelk_cbk] vol-server: 993914: INODELK (null) (--) ==> -1 (No such file or directory)
[2013-03-01 11:34:28.040435] I [server3_1-fops.c:1338:server_flush_cbk] vol-server: 993938: FLUSH -2 (--) ==> -1 (No such file or directory)

Solution:
Moved them to DEBUG log level if error number equlas to ENOENT else to ERROR log level.
It will help in decreasing the size of log files at INFO log level.
For server_open_cbk and for some other functions we already have a patch, below is the URL for it.
URL- http://review.gluster.org/6241
This patch solves logging problem for functions server_inodelk_cbk and server_flush_cbk.

Change-Id: I57372e851371e466f1674726015e28378b826f5f
BUG: 1029372
Signed-off-by: Vikhyat Umrao
Reviewed-on: http://review.gluster.org/6252
Reviewed-by: Niels de Vos 
Tested-by: Gluster Build System

server/rpc: Numerous entries of error - "No such file or directory" in bricks log

2013-11-11T12:36:45+00:00

Problem:
Messages were getting logged very frequently at log level INFO.

One of the log file snippet -

[2013-10-27 00:05:01.501355] I [server3_1-fops.c:1707:server_stat_cbk] 0-vol-server: 24846575: STAT (null) (--) ==> -1 (No such file or directory)
[2013-10-27 00:05:01.505101] I [server3_1-fops.c:1707:server_stat_cbk] 0-vol-server: 24846577: STAT (null) (--) ==> -1 (No such file or directory)
[2013-10-27 00:05:01.507299] I [server3_1-fops.c:1707:server_stat_cbk] 0-vol-server: 24846578: STAT (null) (--) ==> -1 (No such file or directory)
[2013-10-20 19:50:35.554563] I [server3_1-fops.c:1538:server_open_cbk] 0-vol-server: 18714687: OPEN 
(01c70ca0-1952-4e82-abee-a07205757d8e) ==> -1 (No such file or directory)
[2013-10-20 19:50:35.555520] I [server3_1-fops.c:1538:server_open_cbk] 0-vol-server: 18714697: OPEN 
(01c70ca0-1952-4e82-abee-a07205757d8e) ==> -1 (No such file or directory)
[2013-10-20 19:50:35.558292] I [server3_1-fops.c:1538:server_open_cbk] 0-vol-server: 18714712: OPEN (null) (--) ==> -1 (No such file or directory)

Solution:
Moved them to DEBUG log level if error number equlas to ENOENT else to ERROR log level.
It will help in decreasing the size of log files at INFO log level.

Change-Id: I23d74320c9c21bbce4805c20295556cc2cc0a8d8
BUG: 808073
Signed-off-by: Vikhyat Umrao 
Reviewed-on: http://review.gluster.org/6241
Reviewed-by: Niels de Vos 
Tested-by: Gluster Build System 
Reviewed-by: Vijay Bellur

glusterfs: zerofill support

2013-11-11T05:25:49+00:00

Add support for a new ZEROFILL fop. Zerofill writes zeroes to a file in
the specified range. This fop will be useful when a whole file needs to
be initialized with zero (could be useful for zero filled VM disk image
provisioning or  during scrubbing of VM disk images).

Client/application can issue this FOP for zeroing out. Gluster server
will zero out required range of bytes ie server offloaded zeroing. In
the absence of this fop,  client/application has to repetitively issue
write (zero) fop to the server, which is very inefficient method because
of the overheads involved in RPC calls  and acknowledgements.

WRITESAME is a  SCSI T10 command that takes a block of data as input and
writes the same data to other blocks and this write is handled
completely within the storage and hence is known as offload . Linux ,now
has support for SCSI WRITESAME command which is exposed to the user in
the form of BLKZEROOUT ioctl.  BD Xlator can exploit BLKZEROOUT ioctl to
implement this fop. Thus zeroing out operations can be completely
offloaded to the storage device , making it highly efficient.

The fop takes two arguments offset and size. It zeroes out 'size' number
of bytes in an opened file starting from 'offset' position.

This patch adds zerofill support to the following areas:
	- libglusterfs
	- io-stats
	- performance/md-cache,open-behind
	- quota
	- cluster/afr,dht,stripe
	- rpc/xdr
	- protocol/client,server
	- io-threads
	- marker
	- storage/posix
	- libgfapi

Client applications can exloit this fop by using glfs_zerofill introduced in
libgfapi.FUSE support to this fop has not been added as there is no system call
for this fop.

Changes from previous version 3:
* Removed redundant memory failure log messages

Changes from previous version 2:
* Rebased and fixed build error

Changes from previous version 1:
* Rebased for latest master

TODO :
     * Add zerofill support to trace xlator
     * Expose zerofill capability as part of gluster volume info

Here is a performance comparison of server offloaded zeofill vs zeroing
out using repeated writes.

[root@llmvm02 remote]# time ./offloaded aakash-test log 20

real	3m34.155s
user	0m0.018s
sys	0m0.040s
[root@llmvm02 remote]# time ./manually aakash-test log 20

real	4m23.043s
user	0m2.197s
sys	0m14.457s
[root@llmvm02 remote]# time ./offloaded aakash-test log 25;

real	4m28.363s
user	0m0.021s
sys	0m0.025s
[root@llmvm02 remote]# time ./manually aakash-test log 25

real	5m34.278s
user	0m2.957s
sys	0m18.808s

The argument log is a file which we want to set for logging purpose and
the third argument is size in GB .

As we can see there is a performance improvement of around 20% with this
fop.

Change-Id: I081159f5f7edde0ddb78169fb4c21c776ec91a18
BUG: 1028673
Signed-off-by: Aakash Lal Das 
Signed-off-by: M. Mohan Kumar 
Reviewed-on: http://review.gluster.org/5327
Tested-by: Gluster Build System 
Reviewed-by: Vijay Bellur

client_t: phase 2, refactor server_ctx and locks_ctx out

2013-10-31T16:32:50+00:00

remove server_ctx and locks_ctx from client_ctx directly and store as
into discrete entities in the scratch_ctx

hooking up dump will be in phase 3

BUG: 849630
Change-Id: I94cea328326db236cdfdf306cb381e4d58f58d4c
Signed-off-by: Kaleb S. KEITHLEY 
Reviewed-on: http://review.gluster.org/5678
Tested-by: Gluster Build System 
Reviewed-by: Anand Avati

libglusterfs: Add monotonic clocking counter for timer thread

2013-10-15T07:14:57+00:00

gettimeofday() returns the current wall clock time and timezone.
Using these functions in order to measure the passage of time
(how long an operation took) therefore seems like a no-brainer.

This time suffer's from some limitations:

a. They have a low resolution: “High-performance” timing by
definition, requires clock resolutions into the microseconds
or better.

b. They can jump forwards and backwards in time: Computer
clocks all tick at slightly different rates, which causes
the time to drift. Most systems have NTP enabled which
periodically adjusts the system clock to keep them in sync
with “actual” time. The adjustment can cause the clock to
suddenly jump forward (artificially inflating your timing
numbers) or jump backwards (causing your timing calculations
to go negative or hugely positive). In such cases timer
thread could go into an infinite loop.

From 'man gettimeofday':
----------
..
..
The time returned by gettimeofday() is affected by discontinuous
jumps in the system time (e.g., if the system administrator manually
changes the system time).  If you need a monotonically increasing
clock, see clock_gettime(2).
..
..
----------

Rationale:

For calculating interval timing for Timer thread, all that’s
needed should be clock as a simple counter that increments
at a stable rate.

This is necessary to avoid the jumps which are caused by using
"wall time", this counter must be monotonic that can never
“tick” backwards, ever.

Change-Id: I701d31e71a85a73d21a6c5cd15583e7a5a645eeb
BUG: 1017993
Signed-off-by: Harshavardhana 
Reviewed-on: http://review.gluster.org/6070
Tested-by: Gluster Build System 
Reviewed-by: Anand Avati

protocol/server: eliminate an unnecessary frame create and destroy

2013-10-02T23:57:22+00:00

connection_cleanup creates a frame solely so that it can be copied.

But the process of copying creates a new frame, and there's nothing
apparently magic about the one being copied; we can eliminate an
unnecessary create and destroy.

Which in the grand scheme of things probably isn't the worst thing
we do, but it's low hanging fruit.

Change-Id: I4a23b84a53e086137b7d4167ad8c20b673d1ffe5
Signed-off-by: Kaleb S. KEITHLEY 
Reviewed-on: http://review.gluster.org/6019
Reviewed-by: Amar Tumballi 
Tested-by: Gluster Build System 
Reviewed-by: Anand Avati

rpc/rpc-lib: rpcsvc should reply when rpc_err is set

2013-09-11T19:32:33+00:00

Problem:
When requests are received on a connection before setvolume
is done, creating frame from the requests fail because there is
no association of the transport with the conn(i.e. xl_private).
xl_private is set only on set_volume. In such cases error response
is not sent from server xlator to that request because of which
operations on mount point are hanging.

Fix:
Set actor return value to RPCSVC_ACTOR_ERROR so that response is sent
even in these cases.

Change-Id: I74d7bc6849fde6c734008d67c1f4bc9d9f7a84f9
BUG: 1006367
Signed-off-by: Pranith Kumar K 
Reviewed-on: http://review.gluster.org/5892
Tested-by: Gluster Build System 
Reviewed-by: Anand Avati

features/locks : Improves debuggability of inode/entry locks.

2013-09-03T13:50:40+00:00

Prints, in the statedump, the information about the mount that
performed the inode/entry lk.

For the entrylks that are granted after a blocked state, the
blocked time is not printed. A patch for that will be sent
later.

Change-Id: Ib0c1ed21fa9328b435f96b590dd343f59814a08d
BUG: 915629
Signed-off-by: Anuradha Talur 
Reviewed-on: http://review.gluster.org/5712
Reviewed-by: Pranith Kumar Karampuri 
Tested-by: Gluster Build System 
Reviewed-by: Vijay Bellur

libglusterfs/client_t client_t implementation, phase 1

2013-07-30T05:22:21+00:00

Implementation of client_t

The feature page for client_t is at
http://www.gluster.org/community/documentation/index.php/Planning34/client_t

In addition to adding libglusterfs/client_t.[ch] it also extracts/moves
the locktable functionality from xlators/protocol/server to libglusterfs,
where it is used; thus it may now be shared by other xlators too.

This patch is large as it is. Hooking up the state dump is left to do
in phase 2 of this patch set.

(N.B. this change/patch-set supercedes previous change 3689, which was
corrupted during a rebase. That change will be abandoned.)

BUG: 849630
Change-Id: I1433743190630a6d8119a72b81439c0c4c990340
Signed-off-by: Kaleb S. KEITHLEY 
Reviewed-on: http://review.gluster.org/3957
Tested-by: Gluster Build System 
Reviewed-by: Amar Tumballi