RHBA-2012:1253 - Bug Fix Advisory

Topic

An updated glusterfs, glusterfs-debuginfo, glusterfs-devel, glusterfs-fuse,
glusterfs-geo-replication, glusterfs-rdma, gluster-swift, vdsm, and
glusterfs-server packages that fix several bugs are now available for Red Hat
Storage 2.0.

Description

Red Hat Storage is software only, scale-out storage that provides flexible and
affordable unstructured data storage for the enterprise.

This update fixes the following bugs:

In certain parallel setxattr workloads, a mismatch in extended attributes of a
directory between different bricks of a volume is noticed. This results in
getxattr returning inconsistent attributes. This bug has been fixed in this
update (BZ#838197).

Appending data continuously to a file on a FUSE mount reports an Input/Output
error after sometime. This has been fixed in this update (BZ#840394).

Writes to a volume fails even when the quota limit has not been reached. This
issues is due to an optimization in the file system. This bug has been fixed in
this update (BZ#840406).

In a CTDB failover setup, when one server is brought down, there is a long
delay (between 6 to 14 minutes) before the IP automatically fails over to the
other server. This issue is due to a bug in the ping timer logic. This bug has
been fixed in this update (BZ#840655).

When a striped volume is created with XFS bricks, wastage of space is noticed.
This issue is due to the speculative preallocation logic in XFS which has been
fixed in this update (BZ#840809).

After creating a stripe-replicate or a distribute-stripe-replicate volume, write
operation fails with an "Invalid argument" error when writing to a file for the
first time. This issue is due to the changes in stripe-coalesce feature which
has been fixed in this update (BZ#842752).

Renaming a directory fails on a striped volume. This issue has been fixed in
this update (BZ#842953).

In a gluster volume mounted using the glusterfs-fuse option, a non-root user
cannot run an executable which has execute permissions. This issue has been
fixed in this update (BZ#843960).

If gluster volume status command is executed and glusterfsd does not respond in
30 minutes with the status,the cluster lock held by the source glusterd is not
released. This results in failure of other operations. This issue has been fixed
in this update (BZ#844803).

Debuggability of statedump has been enhanced in this update. The enhancements
includes start and end marker addition, using pthread_mutex_trylock instead of
pthread_mutex_lock to avoid deadlocks, appending timestamp to names of the
statedump files, etc (BZ#844802).

The self-heal daemon generates excessive amount of logs. These logs results in
large disk space consumption. This issue has been fixed by minimizing the logs
in this update (BZ#844804).

The VDSM module of Red Hat Storage throws an error during boot while trying to
create multipath devices on local storage due to the default multipath
configuration file. This issue has been fixed in this update (BZ#837869).

When a multipath vector is resized to a smaller size, device-mapper-multipath
does not reassign the pointer to an array. If the array location changes by
reducing its size, device-mapper-multipath corrupts its memory. This issue has
been fixed in this update (BZ#845429).

Some remote procedure call (RPC) handlers of glusterd and glusterfsd return a
value of -1 even when a reply may have been submitted to the remote caller. This
will potentially result in double free causing glusterd to crash. The RPC
handler functions have been modified to return success for RPC layer once reply
is submitted. This issue has been fixed in this update (BZ#851109 and
BZ#851410).

Solution

Before applying this update, make sure all previously-released errata
relevant to your system have been applied.

This update is available via the Red Hat Network. Details on how to
use the Red Hat Network to apply this update are available at
https://access.redhat.com/knowledge/articles/11258

Affected Products

• Red Hat Enterprise Linux Server 6 x86_64
• Red Hat Enterprise Linux Server 5 x86_64
• Red Hat Gluster Storage Server for On-premise 2.0 x86_64
• Red Hat Storage for Public Cloud (via RHUI) 2.0 x86_64

Fixes

• BZ - 837869 - VDSM copies a multipath.conf that is accepting all devices by default and throwing errors on local disk.
• BZ - 838197 - dht_aggregate 'user.swift.metadata' mismatch
• BZ - 840394 - appending data to a file on fuse mount fails with EIO
• BZ - 840655 - CTDB: Noticeable delay in CTDB failover(ranging between 6-14 mins)
• BZ - 842752 - dd fails with "Invalid argument" error on the mount point while creating the file for first time on stripe with replicate volume
• BZ - 842953 - Renaming of directories fails on stripe volume
• BZ - 843960 - Permission Denied error when executing userfiles from RHS (gluster) volume
• BZ - 844802 - improvements in statedumps
• BZ - 844803 - call_bail of a frame in glusterd might lead to stale locks in the cluster
• BZ - 844804 - excessive logging in self-heal daemon should be reduced
• BZ - 851109 - glusterd core dumps when statedump is taken under heavy load
• BZ - 851410 - glusterfsd core dumps when statedump is taken under heavy load

CVEs

(none)

References

(none)