Chapter 4. Known Issues

This chapter provides a list of known issues at the time of release.


Bugzilla IDs that are not hyperlinked are private bugs that have potentially sensitive data attached.

4.1. Red Hat Gluster Storage

Issues related to glusterd

If the enable-shared-storage option is disabled when any one of the glusterd is down, disabling the shared storage operation will be a success. However, subsequent requests of enabling and disabling of enable-shared-storage operations will fail.
Workaround: Run the following commands to overcome this behavior:
# gluster v delete gluster_shared_storage
# gluster v set all cluster.enable-shared-storage enable
Performing add-brick to increase replica count while I/O is going on can lead to data loss.
Workaround: Ensure that increasing replica count is done offline, i.e. without clients accessing the volume.
On a multi node setup where NFS-Ganesha is configured, if the setup has multiple volumes and a node is rebooted at the same time as when volume is stopped, then, once the node comes up the volume status shows that volume is in started state where as it should have been stopped.
Workaround: Restarting the glusterd instance on the node where the volume status reflects started resolves the issue.
glusterd takes time to initialize if the setup is slow. As a result, by the time /etc/fstab entries are mounted, glusterd on the node is not ready to serve that mount, and the glusterd mount fails. Due to this, shared storage may not get mounted after node reboots.
Workaround: If shared storage is not mounted after the node reboots, check if glusterd is up and mount the shared storage volume manually.
If a node is deleted from the NFS-Ganesha HA cluster without performing umount, and then a peer detach of that node is performed, that volume is still accessible in /var/run/gluster/shared_storage/ location even after removing the node in the HA-Cluster.
Workaround: After a peer is detached from the cluster, manually unmount the shared storage on that peer.

Issues related to gdeploy

Currently in a gdeploy configuration file, the ssl_enable option is part of the volume section. If more than one volume section is used in a single gdeploy configuration file for a single storage pool and ssl_enable is set in all the volume sections, then the SSL operation steps are performed multiple times. This fails to mount the older volumes. Thus, users will not be able to set SSL with a single line in the gdeploy configuration file.
Workaround: If there are more than one volume sections in a single gdeploy configuration file for a single storage pool, set the variable enable_ssl under only one volume section and set the keys: 'client.ssl', value: 'on'; 'server.ssl', value: 'on';'auth.ssl-allow', value: comma separated SSL hostnames

Issues related to Arbitrated Volumes

Currently, if the data bricks of the arbiter volume are completely consumed, further creation of new data entries may succeed in the arbiter brick without failing with an ENOSPC error. However, the clients will correctly receive the creation failure error on the mount point. Thus the arbiter bricks might have more entries. When an rm -rf command is executed from the client, readdir operation is performed on one of the databricks to get the list of files to deleted. Consequently, only those entries will get deleted on all bricks. When the rmdir command is executed on the parent directory, it succeeds on the data bricks but fails on the arbiter with an ENOTEMPTY error because it has some files in it.
Workaround: If the deletion from the mount does not encounter an error while the arbiter bricks still contain the directories, the directories and its associated GFID symlinks need to be manually removed. If the directory to be deleted contains files, these files and their associated GFID hard links need to be removed manually.
If some of the bricks of a replica or arbiter sub volume go offline or get disconnected from the client while a rm -rf command is being executed, the directories may re-appear when the bricks are back online and self-heal is complete. When the user tries to create a directory with the same name from the mount, it may heal this existing directory into other DHT subvolumes of the volume.
Workaround: If the deletion from the mount did not complete, but the bricks still contain the directories, the directories and its associated GFID symlink must be removed manually. If the directory to be deleted contains files, these files and their associated GFID hard links need to be removed manually.
If a file is being created on all bricks but succeeds only on the arbiter brick, the application using the file will fail. But during self-heal, the file gets created on the data bricks with arbiter brick marked as the data source. Since data self-heal should not be done from the arbiter brick, the output for the gluster volume heal volname info command will list the entries indefinitely.
Workaround: If the output of the gluster volume heal volname info command indefinitely displays the pending heals for a file, check if the issue is persistent by performing the following steps:
  1. Use the getfattr command to check the following:
    • If the trusted.afr.volname-client* xattrs are zero on the data bricks
    • If the trusted.afr.volname-client* xattrs is non-zero on the arbiter brick only for the data part. The data part is the first 4 bytes.
      For example:
      #getfattr -d -m . -e hex /bricks/arbiterbrick/file |grep trusted.afr.volname*
      getfattr: Removing leading '/' from absolute path names
  2. If the command output matches the mentioned state, delete the xattr using the following command:
    # for i in $(getfattr -d -m . -e hex /bricks/arbiterbrick/file |grep trusted.afr.volname*|cut -f1 -d'='); do  setfattr -x $i file; done

Issues related to Distribute (DHT) Translator

The automatic file replication (AFR) self-heal can have a partially healed file if the brick containing the AFR self-heal source file goes offline during a heal operation. If this partially healed file is migrated before the brick is back online, the migrated file would have incorrect data and the original file would be deleted.

Issues related to Replication (AFR)

In a replicate volume, if a gluster volume snapshot is created when a file creation is in progress, the file may be present in one brick of the replica but not the other brick on the snapshotted volume. Due to this, when this snapshot is restored and a rm -rf dir command is executed on a directory from the mount, it may fail with an ENOTEMPTY error.
Workaround: If you receive the ENOTEMPTY error during the rm -rf dir command execution, but the output of the ls command of the directory shows no entries, check the backend bricks of the replica to verify if files exist on some bricks and not the other. Execute the stat command with the file name from the mount, so that it is healed to all bricks of the replica. Once the file is completely healed, executing the rm -rf dir command is successful.

Issues related to gNFS

From Red Hat Gluster Storage 3.2 onwards, for every volume the option nfs.disable will be explicitly set to either on or off. The default value for new volumes created is on, due to which these volumes will not be exported via. Gluster NFS. The snapshots which were created from 3.1.x version or earlier does not have this volume option.
Workaround: Execute the following command on the volumes:
# gluster v set nfs.disable volname off
The restored volume will not be exported via. Gluster NFS.

Issues related to Tiering

If the gluster volume tier attach command times out, it could result in either of two situations. Either the volume does not become a tiered volume, or the tier daemon is not started.
Workaround: When the timeout is observed, perform the following:
  1. Check if the volume has become a tiered volume.
    • If not, then rerun the gluster volume tier attach command.
    • If it has, then proceed with the next step.
  2. Check if the tier daemons were created on each server.
    • If the tier daemons were not created, execute the following command:
      # gluster volume tier volname start
Listing the entries on a snapshot of a tiered volume displays incorrect permissions for some files. This is because the User Serviceable Snapshot (USS) returns the stat information for the link to files in the cold tier instead of the actual data file. These files appear to have -----T permissions.
Workaround: FUSE clients can work around this issue by applying any of the following options:
  • use-readdirp=no This is the recommended option.
  • attribute-timeout=0
  • entry-timeout=0
NFS clients can work around the issue by applying the noac option.
When a tier is attached while I/O operation is in progress on an NFS mount, I/O pauses temporarily, usually for between 3 to 5 minutes.
Workaround: If I/O does not resume within 5 minutes, use the gluster volume start volname force command to resume I/O without interruption.
Files with hard links are not promoted or demoted on tiered volumes.
There is no known workaround for this issue.
A race condition between tier migration and hard link creation results in the hard link operation failing with a File exists error, and logging Stale file handle messages on the client. This does not impact functionality, and file access works as expected.
This race occurs when a file is migrated to the cold tier after a hard link has been created on the cold tier, but before a hard link is created to the data on the hot tier. In this situation, the attempt to create a hard link on the hot tier fails. However, because the migration converts the hard link on the cold tier to a data file, and a linkto already exists on the cold tier, the links exist and works as expected.
When hot tier storage is full, write operations such as file creation or new writes to existing files fail with a No space left on device error, instead of redirecting writes or flushing data to cold tier storage.
Workaround: If the hot tier is not completely full, it is possible to work around this issue by waiting for the next CTR promote/demote cycle before continuing with write operations.
If the hot tier does fill completely, administrators can copy a file from the hot tier to a safe location, delete the original file from the hot tier, and wait for demotion to free more space on the hot tier before copying the file back.
Migration from the hot tier fails when the hot tier is completely full because there is no space left to set the extended attribute that triggers migration.
Corrupted files can be identified for promotion and promoted to hot tier storage.
In rare circumstances, corruption can be missed by the BitRot scrubber. This can happen in two ways:
  1. A file is corrupted before its checksum is created, so that the checksum matches the corrupted file, and the BitRot scrubber does not mark the file as corrupted.
  2. A checksum is created for a healthy file, the file becomes corrupted, and the corrupted file is not compared to its checksum before being identified for promotion and promoted to the hot tier, where a new (corrupted) checksum is created.
When tiering is in use, these unidentified corrupted files can be 'heated' and selected for promotion to the hot tier. If a corrupted file is migrated to the hot tier, and the hot tier is not replicated, the corrupted file cannot be accessed or migrated back to the cold tier.
When a User Serviceable Snapshot is enabled, attaching a tier succeeds, but any I/O operations in progress during the attach tier operation may fail with stale file handle errors.
Workaround: Disable User Serviceable Snapshots before performing attach tier. Once attach tier has succeeded, User Serviceable Snapshots can be enabled.

Issues related to Snapshot

If NFS-ganesha was enabled while taking a snapshot, and during the restore of that snapshot it is disabled or shared storage is down, then the snapshot restore will fail.
Snapshot create might fail, if a brick has started but not all translators have initialized.
When a volume is down and there is an attempt to access .snaps directory, a negative cache entry is created in the kernel Virtual File System (VFS) cache for the .snaps directory. After the volume is brought back online, accessing the .snaps directory fails with an ENOENT error because of the negative cache entry.
Workaround: Clear the kernel VFS cache by executing the following command:
# echo 3 > /proc/sys/vm/drop_caches
Note that this can cause temporary performance degradation.
If the User Serviceable Snapshot feature is enabled, and a directory has a pre-existing .snaps folder, then accessing that folder can lead to unexpected behavior.
Workaround: Rename the pre-existing .snaps folder with another name.
Performing operations which involve client graph changes such as volume set operations, restoring snapshot, etc. eventually leads to out of memory scenarios for the client processes that mount the volume.
Performing a snapshot restore while glusterd is not available in a cluster node or a node is unavailable results in the following errors:
  • Executing the gluster volume heal vol-name info command displays the error message Transport endpoint not connected.
  • Error occurs when clients try to connect to glusterd service.
Workaround: Perform snapshot restore only if all the nodes and their corresponding glusterd services are running. Start glusterd by running the following command:
# service glusterd start
On restoring a snapshot which was created while the rename of a directory was in progress ( the directory has been renamed on the hashed sub-volume but not on all of the sub-volumes), both the old and new directories will exist and have the same GFID. This can cause inconsistencies and issues accessing files in those directories.
In DHT, a rename (source, destination) of a directory is done first on the hashed sub-volume and if successful, on the remaining sub-volumes. At this point in time, both source and destination directories are present in the volume with same GFID - destination on hashed sub-volume and source on rest of the sub-volumes. A parallel lookup (on either source or destination) at this time can result in creation of these directories on the sub-volumes on which they do not yet exist- source directory entry on hashed and destination directory entry on the remaining sub-volumes. Hence, there would be two directory entries - source and destination - having the same GFID.
If a node/brick is down, the snapshot create command fails even with the force option. This is an expected behavior.
LUKS encryption over LVM is currently not supported.
User Serviceable Snapshots is not supported on Erasure Coded (EC) volumes.

Issues related to Geo-replication

If a geo-replication session is created while gluster volume rebalance is in progress, then geo-replication may miss some files/directories sync to slave volume. This is caused because of internal movement of files due to rebalance.
Workaround: Do not create a geo-replication session if the master volume rebalance is in progress.
If the quick-read performance feature is enabled on the geo-rep slave volume, it could serve stale data as it fails to invalidate its cache in a corner case. This could affect applications reading slave volume as it might get served with stale data.
Workaround: Disable quick read performance feature on the slave volume:
# gluster vol set slave-vol-name quick-read off
With quick-read performance feature disabled, slave will not serve stale data and serves consistent data.
Geo-replication configuration changes when one or more nodes are down in the Master Cluster. Due to this, the nodes that are down will have the old configuration when the nodes are up.
Workaround: Execute the Geo-replication config command again once all nodes are up. With this, all nodes in Master Cluster will have same Geo-replication config options.
Sync performance for geo-replicated storage is reduced when the master volume is tiered, resulting in slower geo-replication performance on tiered volumes.
During file promotion, the rebalance operation sets the sticky bit and suid/sgid bit. Normally, it removes these bits when the migration is complete. If readdirp is called on a file before migration completes, these bits are not removed and remain applied on the client.
If rsync happens while the bits are applied, the bits remain applied to the file as it is synced to the destination, impairing accessibility on the destination. This can happen in any geo-replicated configuration, but the likelihood increases with tiering as the rebalance process is continuous.
The Geo-replication worker goes to faulty state and restarts when resumed. It works as expected when it is restarted, but takes more time to synchronize compared to resume.
The Changelog History API expects brick path to remain the same for a session. However, on snapshot restore, brick path is changed. This causes the History API to fail and geo-rep to change to Faulty.


  1. After the snapshot restore, ensure the master and slave volumes are stopped.
  2. Backup the htime directory (of master volume).
    cp -a <brick_htime_path> <backup_path>


    Using -a option is important to preserve extended attributes.
    For example:
    cp -a /var/run/gluster/snaps/a4e2c4647cf642f68d0f8259b43494c0/brick0/b0/.glusterfs/changeslogs/htime  /opt/backup_htime/brick0_b0
  3. Run the following command to replace the OLD path in the htime file(s) with the new brick path, where OLD_BRICK_PATH is the brick path of the current volume, and NEW_BRICK_PATH is the brick path after snapshot restore.
    find <new_brick_htime_path> - name 'HTIME.*' -print0  | \
    xargs -0 sed -ci 's|<OLD_BRICK_PATH>|<NEW_BRICK_PATH>|g'
    For example:
    find /var/run/gluster/snaps/a4e2c4647cf642f68d0f8259b43494c0/brick0/b0/.glusterfs/changelogs/htime/ -name 'HTIME.*' -print0  | \
    xargs -0 sed -ci 's|/bricks/brick0/b0/|/var/run/gluster/snaps/a4e2c4647cf642f68d0f8259b43494c0/brick0/b0/|g'
  4. Start the Master and Slave volumes and Geo-replication session on the restored volume. The status should update to Active.

Issues related to Self-heal

When files are accidentally deleted from a brick in a replica pair in the back-end, and gluster volume heal VOLNAME full is run, then there is a chance that the files may not heal.
Workaround: Perform a lookup on the files from the client (mount). This triggers the heal.
If you write to an existing file and go over the _AVAILABLE_BRICK_SPACE_, the write fails with an I/O error.
Workaround: Use the cluster.min-free-disk option. If you routinely write files up to nGB in size, then you can set min-free-disk to an mGB value greater than n.
For example, if your file size is 5GB, which is at the high end of the file size you will be writing, you might consider setting min-free-disk to 8 GB. This ensures that the file will be written to a brick with enough available space (assuming one exists).
# gluster v set _VOL_NAME_ min-free-disk 8GB

Issues related to replace-brick operation

  • After the gluster volume replace-brick VOLNAME Brick New-Brick commit force command is executed, the file system operations on that particular volume, which are in transit, fail.
  • After a replace-brick operation, the stat information is different on the NFS mount and the FUSE mount. This happens due to internal time stamp changes when the replace-brick operation is performed.

Issues related to NFS

  • After you restart the NFS server, the unlock within the grace-period feature may fail and the locks held previously may not be reclaimed.
  • fcntl locking (NFS Lock Manager) does not work over IPv6.
  • You cannot perform NFS mount on a machine on which glusterfs-NFS process is already running unless you use the NFS mount -o nolock option. This is because glusterfs-nfs has already registered NLM port with portmapper.
  • If the NFS client is behind a NAT (Network Address Translation) router or a firewall, the locking behavior is unpredictable. The current implementation of NLM assumes that Network Address Translation of the client's IP does not happen.
  • nfs.mount-udp option is disabled by default. You must enable it to use posix-locks on Solaris when using NFS to mount on a Red Hat Gluster Storage volume.
  • If you enable the nfs.mount-udp option, while mounting a subdirectory (exported using the nfs.export-dir option) on Linux, you must mount using the -o proto=tcp option. UDP is not supported for subdirectory mounts on the GlusterFS-NFS server.
  • For NFS Lock Manager to function properly, you must ensure that all of the servers and clients have resolvable hostnames. That is, servers must be able to resolve client names and clients must be able to resolve server hostnames.

Issues related to NFS-Ganesha

The dbus command used to export the volumes fails, if the volumes are exported before completing nfs-ganesha start up.
Workaround: Restart the nfs-ganesha process and then export the volumes.
In case of NFS-Ganesha, the memory created for a cache entry is recycled instead of freeing it. For example, if there is a file "foo" and it is removed from different client cache entry for "foo", it still exists. As a result, memory used by NFS-Ganesha will increase till cache is full.
While adding a node to an existing Ganesha cluster, the following error messages are displayed, intermittently:
Error: Some nodes had a newer tokens than the local node. Local node's tokens were updated. Please repeat the authentication if needed
Error: Unable to communicate with pcsd
Workaround: These messages can safely be ignored since there is no known functionality impact.
The Corosync service will crash, if ifdown is performed after setting up the ganesha cluster. This may impact the HA functionality.
If a volume is being accessed by heterogeneous clients (i.e, both NFSv3 and NFSv4 clients), it is observed that NFSv4 clients take longer time to recover post virtual-IP failover due to node shutdown.
Workaround: Use different VIPs for different access protocol (i.e, NFSv3 or NFSv4) access.
If gluster volume stop operation on a volume exported via NFS-ganesha server fails, there is a probability that the volume will get unexported on few nodes, inspite of the command failure. This will lead to inconsistent state across the NFS-ganesha cluster.
Workaround: To restore the cluster back to normal state, perform the following
  • Identify the nodes where the volume got unexported
  • Re-export the volume manually using the following dbus command:
    # dbus-send --print-reply --system --dest=org.ganesha.nfsd /org/ganesha/nfsd/ExportMgr org.ganesha.nfsd.exportmgr.AddExport string:/var/run/gluster/shared_storage/nfs-ganesha/exports/export.<volname>.conf string:""EXPORT(Path=/<volname>)"""
When a READDIR is issued on directory which is mutating, the cookie sent as part of request could be of the file already deleted. In such cases, server returns BAD_COOKIE error. Due to this, some applications (like bonnie test-suite) which do not handle such errors may error out.
This is an expected behaviour of NFS server and the applications has to be fixed to fix such errors.
If any of the PCS resources are in the failed state, then the teardown requires a lot of time to complete. Due to this, the command gluster nfs-ganesha disable will timeout.
Workaround: If gluster nfs-ganesha disable encounters a timeout, perform the pcs status and check whether any resource is in failed state. Then perform the cleanup for that resource using following command:
# pcs resource --cleanup <resource id>
Re-execute the gluster nfs-ganesha disable command.
After removing a file, the nfs-ganesha process does a lookup on the removed entry to update the attributes in case of any links present. Due to this, as the file is deleted, lookup will fail with ENOENT resulting in a misleading log message in gfapi.log.
This is an expected behaviour and there is no functionality issue here. The log message needs to be ignored in such cases.
When vdsmd and abrt services are installed alongside each other, vdsmd overwrites abrt core dump configuration in /proc/sys/kernel/core_pattern file. This prevents NFS-Ganesha from generating core dumps.
Workaround: Set core_dump_enable to false in /etc/vdsm/vdsm.conf file to disable core dumps, then restart the abrt-ccpp service:
# systemctl restart abrt-ccpp
nfs-ganesha service monitor script which triggers IP failover runs periodically every 10 seconds. By default, the ping-timeout of the glusterFS server (after which the locks of the unreachable client gets flushed) is 42 seconds. After an IP failover, some locks are not cleaned by the glusterFS server process. Therefore, reclaiming the lock state by NFS clients fails.
Workaround: It is recommended to set the nfs-ganesha service monitor period interval (default 10s) to at least twice the Gluster server ping-timeout (default 42s) period.
Therefore, you must decrease the network ping-timeout by using the following command:
# gluster volume set <volname> <ping_timeout_value>
or increase nfs-service monitor interval time by using the following commands:
# pcs resource op remove nfs-mon monitor
# pcs resource op add nfs-mon monitor interval=<interval_period_value> timeout=<timeout_value>
PCS cluster IP resources may enter FAILED state during failover/failback of VIP in NFS-Ganesha HA cluster. As a result, VIP is inaccessible resulting in mount failures or system freeze.
Workaround: Clean up the failed resource by using the following command:
# pcs resource cleanup resource-id
After a reboot, systemd may interpret NFS-Ganesha to be in STARTED state when it is not running.
Workaround: Manually start the NFS-Ganesha process.
Executing the gluster nfs-ganesha disable command stops the NFS-Ganesha service. In case of pre-exported entries, NFS-Ganesha may enter FAILED state.
Workaround: Restart the NFS-Ganesha process after failure and re-run the following command:
# gluster nfs-ganesha disable

Issues related to Object Store

  • The GET and PUT commands fail on large files while using Unified File and Object Storage.
    Workaround: You must set the node_timeout=60 variable in the proxy, container, and the object server configuration files.

Issues related to Red Hat Gluster Storage Volumes

Large number of inodes can cause the itable to be locked for a longer period during inode status dump. This behavior causes performance issues on clients command timeout on inode status dump.
Workaround: Reduce the LRU inodes when performing ‘inode status’ by running the following commands.
Set to small value:
# gluster v set v1 inode-lru-limit 256
Take inode dump:
# gluster v status v1 inode
Set to previous value:
# gluster v set v1 inode-lru-limit 16384
On a volume, when read and write operations are in progress and simultaneously a rebalance operation is performed followed by a remove-brick operation on that volume, then the rm -rf command fails on a few files.
When a brick process dies, BitD tries to read from the socket used to communicate with the corresponding brick. If it fails, BitD logs the failure to the log file. This results in many messages in the log files, leading to the failure of reading from the socket and an increase in the size of the log file.
Due to an unhandled race in the RPC interaction layer, brick down notifications may result in corrupted data structures being accessed. This can lead to NULL pointer access and segfault.
Workaround: When the Bitrot daemon (bitd) crashes (segfault), you can use volume start VOLNAME force to restart bitd on the node(s) where it crashed.
A successful scrub of the filesystem (objects) is required to see if a given object is clean or corrupted. When a file is corrupted and a scrub has not been run on the filesystem, there is a good chance of replicating corrupted objects in cases when the brick holding the good copy was offline when I/O was performed.
Workaround: Objects need to be checked on demand for corruption during healing.

Issues related to Samba

Snapshot volumes are read-only. All snapshots are made available as directories inside .snaps directory. Even though snapshots are read-only, the directory attribute of snapshots is same as the directory attribute of root of snapshot volume, which can be read-write. This can lead to confusion, because Windows will assume that the snapshots directory is read-write. Restore previous version option in file properties gives open option. It will open the file from the corresponding snapshot. If opening of the file also creates temp files (for example, Microsoft Word files), the open will fail. This is because temp file creation will fail because snapshot volume is read-only.
Workaround: Copy such files to a different location instead of directly opening them.
When vdsm and abrt's ccpp addon are installed alongside each other, vdsmd overwrites abrt's core dump configuration in /proc/sys/kernel/core_pattern. This prevents Samba from generating core dumps due to SELinux search denial on new coredump location set by vdsmd.
Workaround: To workaround this issue, execute the following steps:
  1. Disable core dumps in /etc/vdsm/vdsm.conf:
    core_dump_enable = false
  2. Restart the abrt-ccpp and smb services:
    # systemctl restart abrt-ccpp
    # systemctl restart smb
Due to a bug in the Linux CIFS client, SMB2.0+ connections from Linux to Red Hat Gluster Storage currently will not work properly. SMB1 connections from Linux to Red Hat Gluster Storage, and all connections with supported protocols from Windows continue to work.
Workaround: If practical, restrict Linux CIFS mounts to SMB version 1. The simplest way to do this is to not specify the vers mount option, since the default setting is to use only SMB version 1. If restricting Linux CIFS mounts to SMB1 is not practical, disable asynchronous I/O in Samba by setting aio read size to 0 in smb.conf file. Disabling asynchronous I/O may have performance impact on other clients
Attempting to upgrade to ctdb version 4 fails when ctdb2.5-debuginfo is installed, because the ctdb2.5-debuginfo package currently conflicts with the samba-debuginfo package.
Workaround: Manually remove the ctdb2.5-debuginfo package before upgrading to ctdb version 4. If necessary, install samba-debuginfo after the upgrade.
Any changes performed by an administrator in a Gluster volume's share section of smb.conf are replaced with the default Gluster hook scripts settings when the volume is restarted.
Workaround: The administrator must perform the changes again on all nodes after the volume restarts.

Issues related to SELinux

Red Hat Gluster Storage does not currently support SELinux Labeled mounts.
On a FUSE mount, SELinux cannot currently distinguish file systems by subtype, and therefore cannot distinguish between different FUSE file systems (BZ#1291606). This means that a client-specific policy for Red Hat Gluster Storage cannot be defined, and SELinux cannot safely translate client-side extended attributes for files tracked by Red Hat Gluster Storage.
A workaround is in progress for NFS-Ganesha mounts as part of BZ#1269584. When complete, BZ#1269584 will enable Red Hat Gluster Storage support for NFS version 4.2, including SELinux Labeled support.
BZ#1291194 , BZ#1292783
Current SELinux policy prevents ctdb's 49.winbind event script from executing smbcontrol. This can create inconsistent state in winbind, because when a public IP address is moved away from a node, winbind fails to drop connections made through that IP address.

Issues related to Sharding

BZ#1520882 , BZ#1568758
When large number of shards are deleted in a large file, the shard translator synchronously sends unlink operation on all the shards in parallel. This action causes replicate translator to acquire locks on the .shard directory in parallel.
After a short period, large number of locks get accumulated in the locks translatory, and the search for possible matching locks is slowed down, sometimes taking several minutes to complete. This behaviour causes timeouts leading to disconnects and also subsequent failure of file deletion leading to stale shards being left out under the .shard directory.
Workaround: Use shard block size of 64 MB as the lower the shard-block-size, the higher the chances of timeouts.
Sharding relies on block count difference before and after every write as gotten from the underlying file system and adds that to the existing block count of a sharded file. But XFS' speculative preallocation of blocks causes this accounting to go bad as it may so happen that with speculative preallocation the block count of the shards after the write projected by xfs could be greater than the number of blocks actually written to.
Due to this, the block-count of a sharded file might sometimes be projected to be higher than the actual number of blocks consumed on disk. As a result, commands like du -sh might report higher size than the actual number of physical blocks used by the file.

General issues

GFID mismatches cause errors
If files and directories have different GFIDs on different back-ends, the glusterFS client may hang or display errors. Contact Red Hat Support for more information on this issue.
The time stamp of files and directories changes on snapshot restore, resulting in a failure to read the appropriate change logs. glusterfind pre fails with the following error: historical changelogs not available. Existing glusterfind sessions fail to work after a snapshot restore.
Workaround: Gather the necessary information from existing glusterfind sessions, remove the sessions, perform a snapshot restore, and then create new glusterfind sessions.
When storage.reserve limits are reached and a directory is created, the directory creation fails with ENOSPC error and lookup on the directory throws ESTALE errors. As a consequence, file operation is not completed.
Workaround: No workaround is available.
Stale statistics cached in the md-cache and readdir-ahead, fail to get updated on write operations from the application. As a result, the application does not see the effect of write operations like size from the statistics which does not reflect the writes that are successfully completed.
Workaround:Turn off performance.stat-prefetch and performance.readdir-ahead options and the application will no longer receive stale statistics.
glusterfind command must be executed from one node of the cluster. If all the nodes of cluster are not added in known_hosts list of the command initiated node, then glusterfind create command hangs.
Workaround: Add all the hosts in peer including local node to known_hosts.
While migrating VMs, libvirt changes the ownership of the guest image, unless it detects that the image is on a shared filesystem and the VMs cannot access the disk images as the required ownership is not available.
Workaround: Before migration, power off the VMs. When migration is complete, restore the ownership of the VM Disk Image (107:107) and start the VMs.
When a replica brick comes back online, it might have self-heal pending. Executing the rm -rf command on the brick will fail with the error message Directory not empty.
Workaround: Retry the operation when there are no pending self-heals.
When the command rm -rf is executed on the parent directory, which has a pending self-heal entry involving purging files from a sink brick, the directory and files awaiting heal may not be removed from the sink brick. Since the readdir for the rm -rf will be served from the source brick, the file pending entry heal is not deleted from the sink brick. Any data or metadata which is pending heal on such a file are displayed in the output of the command heal-info, until the issue is fixed.
Workaround: If the files and parent directory are not present on other bricks, delete them from the sink brick. This ensures that they are no longer listed in the next 'heal-info' output.
Due to incomplete error reporting, statedump is not generated after executing the following command:
# gluster volume statedump volume client host:port
Workaround: Verify that the host:port is correct in the command.
The resulting statedump file(s) are placed in /var/run/gluster on the host running the gfapi application.

Issues related to Red Hat Gluster Storage AMI

The redhat-storage-server package is not installed by default in a Red Hat Gluster Storage Server 3 on Red Hat Enterprise Linux 7 AMI image.
Workaround: It is highly recommended to manually install this package using yum.
# yum install redhat-storage-server
The redhat-storage-server package primarily provides the /etc/redhat-storage-release file, and sets the environment for the storage node. package primarily provides the /etc/redhat-storage-release file, and sets the environment for the storage node.

Issues related to Red Hat Gluster Storage Web Administration

When central store (etcd) is stopped which could happen either due to stopping of etcd or shutting down the Web Administration server node itself, all the Web Administration services start reporting exceptions regarding reachability to the etcd. As a consequence, Web Administration services crash as etcd is not reachable.
Workaround: Once etcd is back, restart Web Administration services.