RHBA-2013:1262 - Bug Fix Advisory

Topic

Updated Red Hat Storage Server 2.1 packages that fix several bugs and add
various enhancements are now available.

Red Hat Storage is a software only scale-out storage solution that provides
flexible and affordable unstructured data storage. Red Hat Storage 2.1 unifies
data storage and infrastructure, increases performance, and improves
availability and manageability to meet enterprise-level storage challenges.

Users of Red Hat Storage are advised to upgrade to these updated packages,
which fix these bugs.

Description

This update fixes the following bugs:

• Previously, when executing the 'ls -al' command on a volume while data is

being PUT using the Swift feature, a temporary directory would be listed for a
short duration. The issue is now fixed. (BZ#839322)

• Previously, the glusterFS process reported memory allocation failures in dmesg

command due to a memory leak. Now, such messages do not appear as the issue is
rectified. (BZ#840811)

• Previously, a 'Permission Denied' error would occur while executing user files

on Red Hat Storage volumes. Now, the access permissions are modified to handle
the error. (BZ#843080)

• Previously, there was excessive logging of redundant GFID-mismatch messages.

Now, the issue is fixed. (BZ#920870)

• Previously, in certain I/O patterns, there was a possibility for the read

system call of an application to not return from the file system, making it
unresponsive. Now,the issue is fixed with this update. (BZ#901649)

• Previously, the order of display of the nodes in the output of 'gluster

rebalance status' command was not consistent. Now, this output displays the
nodes in a consistent manner. (BZ#888390)

• Previously, if a machine went down during a transaction, it would result in a

stale lock. Now, it is fixed. (BZ#858333)

• Previously, the memory used by the Rebalance process was being freed twice

when the process was just about to complete, leading to a crash. Now the memory
issue is fixed. (BZ#856013)

• Previously, using the Red Hat Storage server VM as a template to create VMs in

the Red Hat Enterprise Virtualization setup would result in the failure of the
peer probe command. Now a command is available to reset the UUID of the local
glusterd, enabling a proper peer probe. (BZ#811493)

• Previously, in certain workloads involving setting and getting of extended

attributes, the log files would overflow with failure messages. Now, this issue
is fixed. (BZ#840812)

• Previously, the GlusterFS-native client could not mount a volume to provide

32-bit inodes. Now, a new 32-bit compatibility option for mounting a glusterFS
volume, called enable-ino32 is available. (BZ#850352)

• Previously, the File Descriptor(FD) resolution in glusterFS server process

would not differentiate between the read-only FDs of geo-replication and NFS FDs
resulting in an error. The fix now handles the FD resolution appropriately.
(BZ#876930)

• Previously, if a volume had 'root-squash' option, internal glusterFS processes

such as rebalance would fail on the volume. Now, a volume that has
root-squashing enabled works fine. (BZ#956770)

• Previously, if a back-end storage disk turned unresponsive, it would result in

the applications on the client becoming unresponsive. Now, a newly added health
checker prevents this issue. (BZ#852578)

• Previously, if an extended attribute was not set on the file, the process for

the brick would log the missing extended attribute as an error. Now, the issue
is fixed. (BZ#918773)

• Previously, the IPtable rule was not present on the nodes making the NFS

volumes unmountable. The fix opened a range of port numbers 38465-38468 and port
2049 to make the NFS volumes mountable. (BZ#979148)

• With this update, the Distributed Hash Table is modified to handle GFID

mismatches. (BZ#915992)

• Now, as per RFC 1912, domain addresses can start or end with a digit.

(BZ#862297)

This update also adds the following enhancements:

• A new distributed Geo-replication feature is added that provides high

availability and better performance. It also uses the glusterfs journaling
feature - changelog, to track the changes, and thus reducing the load on the
master. (BZ#850514)

• A new 'gluster pool list' command is added for listing all the storage servers

in the trusted storage pool. (BZ#870940)

• A logging enhancement is made to the management daemon to log the checksum

mismatches between peers of a volume. (BZ#919588)

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/site/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 2.1 x86_64
• Red Hat Storage for Public Cloud (via RHUI) 2.1 x86_64

Fixes

• BZ - 811493 - When creating a VM from a template, peer UUIDs are the same
• BZ - 840812 - dht_aggregate 'user.swift.metadata' mismatch
• BZ - 843080 - Permission Denied error when executing userfiles from RHS (gluster) volume
• BZ - 868877 - [3.3.0.4rhs-34.el6rhs.x86_64.rpm] geo-rep status gives N/A.
• BZ - 971310 - Unhandled Python exception in dumpStorageTable.pyc when running sosreport on RHSC node
• BZ - 979148 - port 2049 not being open for nfs

CVEs

(none)

References

(none)