glusterfs.git/xlators/protocol/server/src/server.c, branch v3.3.0qa30

protocol/server: Handle server send reply failure gracefully.

2012-03-19T15:11:43+00:00

Server send reply failure should not call server connection cleanup because
if a reconnection happens with in the grace-timeout the connection object is
reused. We must cleanup only on grace-timeout.

Change-Id: I7d171a863382646ff392031c2b845fe4f0d3d5dc
BUG: 803365
Signed-off-by: Mohammed Junaid 
Reviewed-on: http://review.gluster.com/2947
Tested-by: Gluster Build System 
Reviewed-by: Vijay Bellur

print bound_xl only once in server statedump

2012-03-19T12:51:28+00:00

since, currently there is only one bound_xl for
all connection objects, it does not make sense
to print the bound_xl for every conn object.

Change-Id: Iaf4a170fe63fb7e80133c449a20f298f0e86364c
BUG: 765495
Signed-off-by: Rahul C S 
Reviewed-on: http://review.gluster.com/2975
Tested-by: Gluster Build System 
Reviewed-by: Vijay Bellur

Logs: Improved logs in lock/unlock execution path

2012-03-18T08:53:41+00:00

Statedump will now start showing the lk-owner of the stack.

Change-Id: I9f650ce9a8b528cd626c8bb595c1bd1050462c86
BUG: 803209
Signed-off-by: Pranith Kumar K 
Reviewed-on: http://review.gluster.com/2968
Tested-by: Gluster Build System 
Reviewed-by: Anand Avati

protocol/server: Clear internal locks on disconnect

2012-03-18T08:09:43+00:00

If there is a disconnect observed on the client when the
inode/entry unlock is issued, but the reconnection to server
happens with in the grace-time period the inode/entry lk will
live and the unlock will never come from that client.
The internal locks should be cleared on disconnect.

Change-Id: Ib45b1035cfe3b1de381ef3b331c930011e7403be
BUG: 803209
Signed-off-by: Pranith Kumar K 
Reviewed-on: http://review.gluster.com/2966
Reviewed-by: Anand Avati 
Tested-by: Gluster Build System

mgmt/glusterd : volume set validation fixes

2012-03-18T07:52:40+00:00

This is the new version of the patch by Kaushik at review.gluster.com/699

The following new option types have been introduced:
 * GF_OPTION_TYPE_INTERNET_ADDRESS_LIST
 * GF_OPTION_TYPE_PRIORITY_LIST
 * GF_OPTION_TYPE_SIZE_LIST
and option types of several options in translators have been updated to use the
new types.

valid_internet_address(), valid_ipv4_address() & valid_ipv6_address() functions
has been updated for * wildcard matching. Previously used standalone wildcard
address checking functions have been removed.

Changes have been done to stripe translator to correctly set, update and use
stripe-blocksize. Also minimum value for block-size has been set to 16KB.

Change-Id: I2aa484ff695f6a915a8fc9a9f965cf0344f41d59
BUG: 765248
Signed-off-by: Kaushal M 
Reviewed-on: http://review.gluster.com/2899
Tested-by: Gluster Build System 
Reviewed-by: Shishir Gowda 
Reviewed-by: Anand Avati

protocol/server: add and remove the transports from the list, inside the lock

2012-03-18T06:27:26+00:00

Till now for graph changes, glusterfs client used to remember the old graph
also. Hence the transport object on the server corresponding the old graph
never received disconnect. But now since the graph cleanup is happening,
transport on the server side gets disconnect for the cleaned up graph.

Server maintains, all the transports in a list. But addition of the new
transport to the list, or removal of the transport from the list is not
happening within the lock. Thus if a thread is accessing a transport
(in cases of statedump, where each transprt's information is dumped),
and the server gets a disconnect on that transport, then it leads to
segfault of the process.

To avoid it do the list (of transports) manipulation inside the lock.

Change-Id: I50e8389d5ec8f1c52b8d401ef8c8ddd262e82548
BUG: 803815
Signed-off-by: Raghavendra Bhat 
Reviewed-on: http://review.gluster.com/2958
Tested-by: Gluster Build System 
Reviewed-by: Anand Avati

protocol/server: Remove connection from conf->conns w.o. race

2012-03-13T11:58:45+00:00

1) Adding the connection to conf->conns used to
happen in conf->mutex, but removing happened under conn->lock.
Fixed that as below.
When the connection object is created conn's ref, bind_ref count
is set to '1'. For bind_ref ref/unref happens under conf->mutex
whenever server_connection_get, put is called.
When bind_ref goes to '0' connection object is removed from
conf->conns under conf->mutex.  After it is removed from the list,
conn_unref is called outside the conf->mutex.
conn_ref/unref still happens under conn->lock.

2) Fixed races in server_connection_cleaup in grace_timer_handler
and server_setvolume.

Change-Id: Ie7b63b10f658af909a11c3327066667f5b7bd114
BUG: 801675
Signed-off-by: Pranith Kumar K 
Reviewed-on: http://review.gluster.com/2911
Tested-by: Gluster Build System 
Reviewed-by: Amar Tumballi 
Reviewed-by: Vijay Bellur

protocol/server: Make conn object ref-counted

2012-03-01T17:12:10+00:00

Change-Id: I992a7f8a75edfe7d75afaa1abe0ad45e8f351c8b
BUG: 796581
Signed-off-by: Pranith Kumar K 
Reviewed-on: http://review.gluster.com/2806
Tested-by: Gluster Build System 
Reviewed-by: Vijay Bellur

transport/socket: configuring tcp window-size

2012-02-29T10:10:44+00:00

Till now, send and recieve buffer window sizes for sockets
were set to a default glusterfs-specific value.
Linux's default window sizes have been found to be better
w.r.t performance, and hence, no more setting it to any
default value.

However, if one wishes, there's the new configuration option:
   network.tcp-window-size 
which takes a size value (int or human readable) and will set
the window size of sockets for both clients and servers.
Nfs clients will also be updated with the same.

Change-Id: I841479bbaea791b01086c42f58401ed297ff16ea
BUG: 795635
Signed-off-by: Rajesh Amaravathi 
Reviewed-on: http://review.gluster.com/2821
Tested-by: Gluster Build System 
Reviewed-by: Amar Tumballi 
Reviewed-by: Jeff Darcy 
Reviewed-by: Vijay Bellur

mempool: adjustments in pool sizes

2012-02-22T12:23:48+00:00

* while creating 'rpc_clnt', the caller knows what would be the ideal
  load on it, so an extra argument to set some pool sizes

* while creating 'rpcsvc', the caller knows what would be the ideal
  load of it, so an extra argument to set request pool size

* cli memory footprint is reduced

Change-Id: Ie245216525b450e3373ef55b654b4cd30741347f
Signed-off-by: Amar Tumballi 
BUG: 765336
Reviewed-on: http://review.gluster.com/2784
Tested-by: Gluster Build System 
Reviewed-by: Vijay Bellur