RHBA-2020:5603 - Bug Fix Advisory

Topic

Updated glusterfs packages that fix several bugs and add various enhancements are now available for Red Hat Enterprise Linux 8.

Description

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

This advisory fixes the following bugs:

• Previously, if the bit-rot feature was disabled and enabled on the same volume it would mark some files as corrupted. Now, the files are signed unconditionally when the bit-rot daemon comes up instead of skipping the files with bit-rot extended attributes. (BZ#1851424)
• Guest access gluster volumes shared using SMB is no longer enabled

by default. (BZ#1775637)

• When file healing failed, gluster retried healing multiple times,

which consumed CPU resources significantly. Continuous healing is now
detected more reliably, and CPU usage is reduced when healing does not
immediately succeed. (BZ#1852736)

• Previously, when IPv6 addresses were used during brick creation, the geo-replication would fail to start causing faulty geo-replication sessions. With this update, geo-replication can parse full brick and hostname from volfile. (BZ#1855966)
• Detecting disconnected hyperconverged hosts took too long and caused

timeouts that resulted in sanlock killing virtual machines when leases
could not be refreshed. Tuning has improved disconnect detection so that
the timeout no longer occurs. (BZ#1845064)

• When the glusterd service crashes, systemd now restarts it automatically

up to 6 times per hour. (BZ#1663557)

• The '--type' option has been added to glusterfind so that either

files (f) or directories (d) or both (both) are shown. Use this option
with the `--full` option to show only the specified file type(s).
(BZ#1383723)

• The 'gluster volume heal $volname info' command no longer takes

blocking locks in order to avoid a possible hang that occurred when a
lock was acquired by a client writing to the same file. (BZ#1721355)

• The replication module now uses eager-lock functionality for fsync

operations, improving performance on write-heavy hyperconverged and
virtualization workloads. (BZ#1838479)

• Split-brain resolution using the command line is now available for

replica 3 volumes. (BZ#1759875)

• Previously, spurious entry heal was triggered even when only the

source bricks were available, which resulted in I/O errors on the
mount when the AFR extended attributes were unintentionally reset.
Entry heals are now triggered only when all the 3 bricks are available,
avoiding GFID split-brain in this situation. (BZ#1804164)

• SELinux policies specific to Red Hat Gluster Storage are now shipped

in Red Hat Gluster Storage packages. (BZ#1460657)

All users of Red Hat Gluster Storage are advised to upgrade to these
updated packages, which resolve these issues and add these enhancements.

Solution

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

For details on how to apply this update, refer to:

https://access.redhat.com/articles/11258

Affected Products

• Red Hat Gluster Storage Server for On-premise 3 for RHEL 8 x86_64
• Red Hat Enterprise Linux for x86_64 8 x86_64

Fixes

• BZ - 1234220 - [Backup]: Glusterfind CLI commands need to verify the accepted names for session/volume, before failing with error(s)
• BZ - 1286171 - Rebalance : Status lists failures on stopping rebalance while it is in progress
• BZ - 1383723 - [GSS][RFE] provide option in glusterfind to display only files as well
• BZ - 1412545 - healing must not change the ctime of the file
• BZ - 1460657 - [RFE] Remove Glusterfs SELinux module from RHEL Distribution selinux-policy, and ship Glusterfs SELinux module as part of glusterfs package
• BZ - 1487177 - change the error message for heal statistics to reflect its supportability for disperse volume
• BZ - 1524457 - check for same hostnames(bricks from same host/node) while adding bricks in erasure (ec) ie disperse volumes and error out accordingly(use force to over-ride)
• BZ - 1529249 - Brick Multiplexing: glusterfsd (brick) process leaks memory on an idle brick multiplex setup
• BZ - 1640573 - [geo-rep]: Transport endpoint not connected with arbiter volumes
• BZ - 1658132 - [Ganesha] Ganesha failed on one node while exporting volumes in loop
• BZ - 1663557 - RFE: systemd should restart glusterd on crash
• BZ - 1667954 - [man page] Gluster(8) - Missing disperse-data parameter Gluster Console Manager man page
• BZ - 1683602 - [Brick-mux] Observing multiple brick processes on node reboot with volume start
• BZ - 1686897 - [CodeChange][Coverity] RETURN_LOCAL in __nlc_inode_ctx_get()
• BZ - 1721355 - Heal Info is hung when I/O is in progress on a gluster block volume
• BZ - 1748865 - [GSS][Duplicate entries in 'ls' output after a volume expansion]
• BZ - 1750211 - [GSS] glustershd logs are filling up with error client-rpc-fops.c:1513:client3_3_inodelk_cbk: remote operation failed [Invalid argument]
• BZ - 1754391 - man page update needed for gluster volume top command
• BZ - 1759875 - afr: support split-brain CLI for replica 3
• BZ - 1761531 - heal not actually healing metadata of a regular file when only time stamps are changed(data heal not required)
• BZ - 1761932 - ctime value is different from atime/mtime on a create of file
• BZ - 1763030 - Set volume option when one of the node is powered off, After powering the node brick processes are offline (with brick-mux enabled)
• BZ - 1763124 - clearly state that replica 2 is deprecated in cli warning when user tries to create replica 2 volume
• BZ - 1763129 - either don't allow or else throw a warning if a user tries to create a volume with replica count greater than 3
• BZ - 1764091 - Setting cluster.heal-timeout requires volume restart
• BZ - 1775637 - Remove guest access by default for GlusterFS volume SMB shares added by hook scripts
• BZ - 1776901 - [GSS] Volume creation and deletion fails in OCS (OCS 3.11.2) (brickmux)
• BZ - 1781550 - glusterfs process memory leak in ior test
• BZ - 1781649 - READDIRP incorrectly updates posix-acl inode ctx
• BZ - 1781710 - default option is disappeared in volume info after volume reset
• BZ - 1783232 - gluster-smb:glusto-test access gluster by cifs test write report Device or resource busy
• BZ - 1784211 - [GSS] - 'gluster volume set <VOLUME> disable.nfs' accidentally killed unexpected process, and forced a data brick offline.
• BZ - 1784415 - Deprecated log rotate command still present
• BZ - 1785577 - gluster processes are having memory leak after enabling SSL
• BZ - 1786516 - on a brickmux setup glusterd consumes very high memory with SSL enabled when compared without SSL enablement (dup of bug 1785577 )
• BZ - 1786681 - quota_fsck script KeyError: 'contri_size'
• BZ - 1787294 - Improve logging in EC, client and lock xlator
• BZ - 1787305 - lots of "Matching lock not found for unlock xxx" when using disperse (ec) xlator
• BZ - 1787310 - [CodeChange] Fix issues reported by Coverity static analysis tool
• BZ - 1787994 - [geo-rep] Help for positional argument SLAVE in schedule_georep.py.in isn't clear.
• BZ - 1790336 - brick crashed when creating and deleting volumes over time (with brick mux enabled only)
• BZ - 1794663 - [brickmux]: glustershd crashed when rebooting 1/3 nodes at regular intervals
• BZ - 1796814 - config ssh-port can accept negative and outside allowed port range value
• BZ - 1804164 - Spurious healing results in heal pending on the volume
• BZ - 1810924 - [SSL]: Memory leak by glusterfsd (issue heal info indefinitely to reproduce) when SSL enabled for Management layer
• BZ - 1812789 - [RHEL-8.1] On distributed-disperse volume, Remove-brick status showing failed on one of the node after few hours of the operation.
• BZ - 1812835 - [RHEL-7]-DH ciphers disabled info flooding glusterd log messages on a SSL enabled setup
• BZ - 1815434 - [samba-disperse] unable to create files in samba share with non-root user for distribute disperse volume
• BZ - 1824790 - Create glusterfs-selinux dist-git repo
• BZ - 1825736 - [GSS] glustershd process crashes intermittently with " uuid_is_null"
• BZ - 1831403 - [RHEL-8.2] On distributed-disperse volume, Remove-brick status showing failed on one of the node after few hours
• BZ - 1836099 - [RHEL 8.2] Rebalance takes long time to migrate the files with holes
• BZ - 1837926 - Snapshot clone fails with wrong error message.
• BZ - 1839137 - FUSE protocol status indication suppressed
• BZ - 1844359 - Place-holder bug for backporting .t fixes from Upstream
• BZ - 1847081 - add-brick functionality is completely broken
• BZ - 1851424 - Bitrot - File is not signed after doing bitrot disable/enable
• BZ - 1851989 - [RHGS 3.5.3]: Regression in IO performance seen in performance test for small file.
• BZ - 1852736 - High CPU usage on EC volume after inservice upgrade of one node in 3 node cluster
• BZ - 1853189 - [RHGS 3.5][NFS-Ganesha][USS].snaps/ directory not readable when accessed over NFS
• BZ - 1854165 - gluster does not release posix lock when multiple glusterfs clients do flock -xo on the same file in parallel
• BZ - 1855966 - [IPV6] Geo-replication session fails to start sync with IPV6 hostnames
• BZ - 1875643 - RHGS3.5.2 for RHEL8 installation media doesn't contain sos package
• BZ - 1875723 - RHGS3.5.2 on RHEL8 installation media doesn't contain dependency packages for pcs package.
• BZ - 1880256 - Geo-rep status going into `Faulty` after the session is started successfully
• BZ - 1881823 - add-brick: Getting an error message while adding a brick from different node to the volume.
• BZ - 1886259 - [Test Day] Command to groupinstall the RH-Gluster-Samba-Server fails
• BZ - 1895301 - [Ganesha][RHEL-8.3] HA cluster status shows "FAILOVER" even all services are running and VIP's are assigned to nodes
• BZ - 1901831 - Rebase redhat-release to RHEL-8.3 for RHGS
• BZ - 1901949 - Product certificates need to be updated in redhat-storage-server for RHEL 8.3

CVEs

(none)

References

(none)