[Gluster-users] Explanation of gluster vol geo-repliation options

Aravinda VK aravinda at kadalu.io
Thu Oct 8 06:37:43 UTC 2020

Hi Felix,

I am adding the details inline. Will open an issue to update the doc. Thanks for reporting.

Let me know if any of these needs more clarity.

> On 08-Oct-2020, at 12:42 AM, Felix Kölzow <felix.koelzow at gmx.de> wrote:
> Dear Community,
> actually, I am looking for a more detailed explaination of all
> geo-replication options which can be obtained by the command:
> gluster volume geo-replication master-vol slave-node::slave-volume config
> I only found
> https://docs.gluster.org/en/latest/Administrator%20Guide/Geo%20Replication/
> but i am wondering if there is a more detailed explanation of all the
> other options listed below.
> Thanks in advance.
> Felix
> access_mount:false
Internal Gluster mounts created by Geo-rep are not accessible from outside. With this option, these mounts can be accessed from outside. Useful for debugging purposes.

> allow_network:
> change_detector:changelog
Use changelogs to detect the changes happened in the Primary Volume. Other available option is xsync which is File system crawl and known limitation to detect deletes.

> change_interval:5
Changelogs rollover interval

> changelog_archive_format:%Y%m
Processed changelogs will be stored in working directory partitioned by the specified format.

> changelog_batch_size:727040
Bytes to limit the changelogs in Batches. Mostly useful if backlog to sync is more.

> changelog_log_file:/var/log/glusterfs/geo-replication/
Changelogs processing log files directory

> changelog_log_level:INFO
Changelogs processing log level

> checkpoint:0
Time will be set here when checkpoint command is run.

> cli_log_file:/var/log/glusterfs/geo-replication/cli.log
CLI Log file path

> cli_log_level:INFO
CLI log level

> connection_timeout:60

> georep_session_working_dir:/var/lib/glusterd/geo-replication
Geo-rep Working dir to store config, status etc

> gfid_conflict_resolution:true
This enables automatic resolution to GFID conflicts in Secondary Volume.

> gluster_cli_options:
Geo-rep runs Gluster CLI when needed, any extra options can be provided using this.

> gluster_command:gluster
If command name is different.

> gluster_command_dir:/usr/sbin
If gluster command is not in PATH(gluster_command_dir=/usr/local/sbin)

> gluster_log_file:/var/log/glusterfs/geo-replication/
Gluster mount(Mounted by Geo-rep process) log directory

> gluster_log_level:INFO
Gluster mount(Mounted by Geo-rep process) log level

> gluster_logdir:/var/log/glusterfs

> gluster_params:aux-gfid-mount acl

> gluster_rundir:/var/run/gluster

> glusterd_workdir:/var/lib/glusterd

> gsyncd_miscdir:/var/lib/misc/gluster/gsyncd

> ignore_deletes:false

> isolated_slaves:

> log_file:/var/log/glusterfs/geo-replication/

> log_level:INFO

> log_rsync_performance:true
After every batches records Rsync performance metrics in log file

> master_disperse_count:1
Not used now

> master_distribution_count:2
Not used now

> master_replica_count:1
Not used now

> max_rsync_retries:10
In case of rsync errors, maximum number of retries

> meta_volume_mnt:/var/run/gluster/shared_storage
If meta volume is used for Active/Passive switching then the mount path.

> pid_file:
Path of Geo-rep worker PID file

> remote_gsyncd:
Path of remote_gsyncd. No need to change this since Geo-rep ssh connection uses proper path prefix in .ssh/authorized_keys file

> replica_failover_interval:1
In case of Active worker goes down, interval to make Passive worker becomes Active.

> rsync_command:rsync
If rsync command is not in PATH(rsync_command=/usr/local/bin/rsync)

> rsync_opt_existing:true
Adds `—existing` to rsync command

> rsync_opt_ignore_missing_args:true
Adds `--ignore-missing-args` to rsync command. Helps to avoid retries in case of file is deleted in Primary Volume at the time of syncing.

> rsync_options:-h
Extra Options to Rsync

> rsync_ssh_options:
Extra options to SSH command

> slave_access_mount:false
Same as access_mount, for debugging purpose

> slave_gluster_command_dir:/usr/sbin
Gluster command directory in remote nodes(Secondary volume)

> slave_gluster_log_file:/var/log/glusterfs/geo-replication-slaves/
Log dir path

> slave_gluster_log_file_mbr:/var/log/glusterfs/geo-replication-slaves
Log directory path in case of non-root Geo-replication

> slave_gluster_log_level:INFO
Log level for Gluster mounts done by Geo-rep process remote side.

> slave_gluster_params:aux-gfid-mount acl
Parameters to Remote mount

> slave_log_file:/var/log/glusterfs/geo-replication-slaves/
Log file for Remote errors

> slave_log_level:INFO
Log level in remote side

> slave_timeout:120
SSH connections timeout

> special_sync_mode:
This is required to set during the failover

> ssh_command:ssh
If ssh command is not in PATH(ssh_command=/usr/local/sbin/ssh)

> ssh_options:-oPasswordAuthentication=no -oStrictHostKeyChecking=no -i
> /var/lib/glusterd/geo-replication/secret.pem
If any additional option to ssh command.

> ssh_options_tar:-oPasswordAuthentication=no -oStrictHostKeyChecking=no
> -i /var/lib/glusterd/geo-replication/tar_ssh.pem
If any additional option to tar + ssh command.

> ssh_port:22
SSH port for all SSH connection between Primary Volume to Secondary volume.

> state_file:/var/lib/glusterd/geo-replication/
Status file directory

> state_socket_unencoded:

> stime_xattr_prefix:

> sync_acls:true
Sync Posix Acls to remote volume.

> sync_jobs:3
Number of parallel rsync/tar processes to sync

> sync_method:rsync

Supported sync methods are rsync and tarssh. rsync - After change detection, uses rsync to sync files. tarssh - After change detection, uses tar + ssh to sync files.

> sync_xattrs:true
Same as use_rsync_xattrs, but it is having a known issue of not working in tarssh mode

> tar_command:tar
In case tar command is not in Path(`tar_command=/usr/local/bin/tar`)

> use_meta_volume:true
In case of Replica bricks only one brick’s worker will be Active and participant in syncing. Other workers will be Passive and becomes active if the previously Active one goes down. With this option Geo-rep maintains a lock file in Meta Volume. Without this option, Geo-rep uses node uuid to detect Active/Passive.

> use_rsync_xattrs:false
Sync user created xattrs to remote volume.

> working_dir:/var/lib/misc/gluster/gsyncd/…..
Not Configurable, Working directory to store the processed changelog files.

> ________
> Community Meeting Calendar:
> Schedule -
> Every 2nd and 4th Tuesday at 14:30 IST / 09:00 UTC
> Bridge: https://bluejeans.com/441850968
> Gluster-users mailing list
> Gluster-users at gluster.org
> https://lists.gluster.org/mailman/listinfo/gluster-users

Aravinda Vishwanathapura

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.gluster.org/pipermail/gluster-users/attachments/20201008/0a53f16f/attachment.html>

More information about the Gluster-users mailing list