glusterfs.git/glusterfsd, branch v3.3.0.5rhs-39

Rebase commit #2

2012-12-12T08:24:00+00:00

Change-Id: Ie983d0b9862cc1401187532ed896e57bd3488e2b
BUG: 871323
Reviewed-on: https://code.engineering.redhat.com/gerrit/1893
Reviewed-by: Vijay Bellur 
Tested-by: Vijay Bellur

fuse: make background queue length configurable

2012-12-12T05:14:14+00:00

* also make 'congestion_threshold' an option
* make 'congestion_threshold' as 75% of background queue length if
  not explicitely specified
* in glusterfsd.c, moved all the fuse option dictionary setting
  code to separate function

Change-Id: Ie1680eefaed9377720770a09222282321bd4132e
Signed-off-by: Amar Tumballi 
BUG: 845214
Reviewed-on: https://code.engineering.redhat.com/gerrit/1860
Reviewed-by: Vijay Bellur 
Tested-by: Vijay Bellur

libglusterfs,mount/fuse: implement gidcache mechanism in fuse-bridge

2012-12-12T05:12:01+00:00

This change genericizes the cache mechanism implemented in commit
8efd2845 into libglusterfs/src/gidcache.[ch] and adds fuse-bridge as
a client. The cache mechanism is fundamentally equivalent, with some
minor changes:

  - Change cache key from uid_t to uint64_t.
  - Modify the cache add logic to locate and use an entry with a
    matching ID, should it already exist. This addresses a bug in
    the existing mechanism where an expired entry supercedes a newly
    added entry in lookup, causing repeated adds and flushing of a
    cache bucket.

The fuse group cache is disabled by default. It can be enabled via
the 'gid-timeout' fuse-bridge translator option and accompanying
mount option (i.e., '-o gid-timeout=1' for a 1s entry timeout).

BUG: 800892
Change-Id: I0b34a2263ca48dbb154790a4a44fc70b733e9114
Signed-off-by: Brian Foster 
Reviewed-on: https://code.engineering.redhat.com/gerrit/1853
Reviewed-by: Vijay Bellur 
Tested-by: Vijay Bellur

mount/fuse: add mount-option "enable-ino32" for the native client

2012-11-19T09:03:08+00:00

From upstream commit c13823bd16b26bc471d3efb15f63b76fbfdf0309:

> By default the GlusterFS-native client uses 64-bit inodes. Some 32-bit
> applications can not handle these correctly. Introduce a client-side
> mount option "enable-ino32" which causes the FUSE-client to squash the
> 64-bit inodes into a 32-bit value.
>
> Change-Id: I3296d16528bfb50457b9675f6b8701234ed82ff0
> BUG: 850352
> Signed-off-by: Niels de Vos 
> Reviewed-on: http://review.gluster.org/3885
> Tested-by: Gluster Build System 
> Reviewed-by: Anand Avati 
> Signed-off-by: Niels de Vos 

BUG: 876679
Change-Id: Ie52ebc7d08f69a11cfa892a057bc9faf71446dcc
Signed-off-by: Niels de Vos 
Reviewed-on: https://code.engineering.redhat.com/gerrit/1585
Reviewed-by: Vijay Bellur 
Tested-by: Vijay Bellur

glusterfsd: rpcsvc actors must return success after an attempt to submit reply

2012-09-04T05:01:44+00:00

rpcsvc attempts to send "error reply" using the req object. If actor has
already performed rpcsvc_submit_generic, then req is destroyed. So if the
actor returned -1 (RPCSVC_ACTOR_ERROR) on failing to submit reply, then req
would be 'free'd' twice and will result in a crash eventually.

Change-Id: I5eae19570202bbe5e154e9cb03390cfeb9b5f223
BUG: 851410
Signed-off-by: Krishnan Parthasarathi 
Reviewed-on: http://review.gluster.org/3863
Tested-by: Gluster Build System 
Reviewed-by: Anand Avati 
Reviewed-on: http://review.gluster.org/3875
Reviewed-by: Vijay Bellur

rpc: Reduce frame-timeout for glusterd connections

2012-08-17T08:16:19+00:00

Reduce frame-timeout for glusterd connections from 30mins to 10 mins. 30mins is
too long when compared to cli timeout of 2mins. Changing to 10mins reduces the
disparity between cli and glusterd.

Also, fix glusterfs_submit_reply() so that a reply is sent even if serialize
failed.

BUG: 843003
Change-Id: Ie8d5ec16fbbb54318a5935a47065e66fd3338b87
Signed-off-by: Kaushal M 
Reviewed-on: http://review.gluster.com/3812
Tested-by: Gluster Build System 
Reviewed-by: Vijay Bellur

fuse: make SELinux support configurable

2012-05-30T05:58:07+00:00

Make support for SELinux labels (extended attributes) configurable
and disabled by default as it can cause significant performance
penalty when enabled (it need not be enabled unless specially crafted
policies are set -- which is not by default)

Change-Id: I97bc4b1c26cf055fd520e9bf2d49e52b14fe7515
BUG: 811217
Signed-off-by: Anand Avati 
Reviewed-on: http://review.gluster.com/3485
Tested-by: Gluster Build System

fuse, glusterfsd: mount logic fixes

2012-05-27T16:55:47+00:00

Commit 7d0397c2 introduced two issues:

 i) broke the libfuse derived mount logic (details below)
ii) in case of a daemonized glusterfs client is ran as daemon, parent
    process can return earlier than the mount is in place, which breaks
    agents that programmatically do a gluster mount via a direct call to
    glusterfs (ie. not via mount(8)).

This patch fixes these issues by a refactor that merges the approaches
sported by commits

  7d0397c2 fuse: allow requests during mount (needed for SELinux labels)
  c5d781e0 upon daemonizing, wait on mtab update to terminate in parent

Original daemonized libfuse event flow is as follows:

  try:
    fd = open("/dev/fuse")
    mount("-oopts,fd=%s" % fd ...)
    mount(8) -f    # manipulate mtab
  except:
    sp = socketpair()
    env _FUSE_COMMFD=sp fusermount -oopts
    fd = receive_fd(sp)

  where fusermount(1) does:
    fd = open("/dev/fuse")
    mount("-oopts,fd=%d" % fd ...)
    sp = atoi(getenv("_FUSE_COMMFD"))
    send_fd(sp, fd)

  daemonize(
    # in child
    fuse_loop(fd)
  )
  # in parent
  exit()

As of 013850c9 (instead of adopting FUSE's 47e61004¹), we went for async
mtab manipulation, and as of c5d781e0, still wanted keep that in sync
with termination of daemon parent, so we changed it to:

  try:
    fd = open("/dev/fuse")
    mount("-oopts,fd=%s" % fd ...)
    pid = fork(
      # in child
      mount(8) -f
    )
  except:
    sp = socketpair()
    env _FUSE_COMMFD=sp fusermount -oopts
    fd = receive_fd(sp)

  daemonize(
    fuse_loop(fd)
  )
  waitpid(pid)
  exit()

(Note the new approch came only to direct [privileged] mount, so fusermount
based mounting was already partially broken.)

As of 7d0397c2, with the purpose of facilitating async mount, the event flow
was practically reduced to:

  fd = open("/dev/fuse")
  fork(
    mount("-oopts,fd=%s" % fd ...)
    fork(
      mount(8) -n
    )
  )

  daemonize(
    fuse_loop(fd)
  )
  exit()

Thus fusermount based mounting become defunct; however, the dead
code was still kept around. So, we should either drop it or fix
it. Also, the mtab manipulator is forked into yet another child
with no purpose, while syncing with it in daemon parent is broken.
mount(2) is neither synced with parent.

Now we are coming to the following scheme:

  fd = open("/dev/fuse")
  pid = fork(
    try:
      mount("-oopts,fd=%s" % fd ...)
      mount(8) -n
    except:
      env _FUSE_DEVFD=fd fusermount -oopts
  )

  where fusermount(1) does:
    fd = getenv("_FUSE_DEVFD")
    mount("-oopts,fd=%s" % fd ...)

  daemonize(
    fuse_loop(fd)
  )
  waitpid(pid)
  exit()

Nb.:

- We can't help losing compatibility with upstream fusermount,
  as it sends back the fd only when mount(2) is completed,
  thus defeating the async mount approach. The
  'getenv("_FUSE_DEVFD")' mechanism is specfic to glusterfs'
  fusermount (at the moment -- sure we can talk about it with
  upstream)

- fusermount opens /dev/fuse at same privilege level as of
  original process², so we can bravely go on with doing the open
  unconditionally in original process

- Original mounting code actually tries to mount through
  fusermount _twice_: if first attempt fails, then, assuming
  subtype support is missing in kernel, it tries again subtype
  stripped. However, this is redundant, as fusermount internally
  also performs the subtype check³. Therefore we simplified the
  logic to have just a single fusermount call.

- we revert the changes to mount.glusterfs as of 7d0397c2, as
  now there is no issue with glusterfs to work around in that scope

¹ http://fuse.git.sourceforge.net/git/gitweb.cgi?p=fuse/fuse;a=blobdiff;f=ChangeLog;h=47e61004;hb=4c3d9b19;hpb=e61b775a
² http://fuse.git.sourceforge.net/git/gitweb.cgi?p=fuse/fuse;a=blob;f=util/fusermount.c;h=b2e87d95#l1023
³ http://fuse.git.sourceforge.net/git/gitweb.cgi?p=fuse/fuse;a=blob;f=util/fusermount.c;h=b2e87d95#l839

Change-Id: I0c4ab70e0c5ad7b27337228749b266bcd0ba941d
BUG: 811217
Signed-off-by: Csaba Henk 
Reviewed-on: http://review.gluster.com/3428
Tested-by: Gluster Build System 
Reviewed-by: Vijay Bellur

glusterfs/rebalance: Register cbk for glusterfs_rebalance_event_notify

2012-05-23T09:02:10+00:00

In case of n/w failures, cbk needs to be passed to
mgmt_submit_request

BUG: 822086
Change-Id: Ie443902b94a09e11a2696b89de44a11fb477ca3e
Signed-off-by: shishir gowda 
Reviewed-on: http://review.gluster.com/3403
Tested-by: Gluster Build System 
Reviewed-by: Amar Tumballi 
Reviewed-by: Raghavendra Bhat 
Reviewed-by: Vijay Bellur

glusterfsd: Make sure mountpoint is an absolute path

2012-05-09T00:00:21+00:00

If the mountpoint path given to glusterfs is not an absolute path, convert it to
an absolute path by concatenating it with the curren working directory.
This prevents cases, where in gluster cannot perform clean unmounts when mount
is done with a relative path.

Change-Id: Ie25add4e1dc59171e522c4244c79a6c148844ab3
BUG: 819466
Signed-off-by: Kaushal M 
Reviewed-on: http://review.gluster.com/3302
Tested-by: Gluster Build System 
Reviewed-by: Amar Tumballi 
Reviewed-by: Anand Avati