Bug 1660225
Summary: | geo-rep does not replicate mv or rename of file | ||||||||||
---|---|---|---|---|---|---|---|---|---|---|---|
Product: | [Community] GlusterFS | Reporter: | perplexed767 <perplexed767> | ||||||||
Component: | geo-replication | Assignee: | Sunny Kumar <sunkumar> | ||||||||
Status: | CLOSED NEXTRELEASE | QA Contact: | |||||||||
Severity: | high | Docs Contact: | |||||||||
Priority: | unspecified | ||||||||||
Version: | 4.1 | CC: | asender, ben, bugs, hgowtham, khiremat, pasik, perplexed767, sunkumar | ||||||||
Target Milestone: | --- | ||||||||||
Target Release: | --- | ||||||||||
Hardware: | aarch64 | ||||||||||
OS: | Linux | ||||||||||
Whiteboard: | |||||||||||
Fixed In Version: | Doc Type: | If docs needed, set a value | |||||||||
Doc Text: | Story Points: | --- | |||||||||
Clone Of: | Environment: | ||||||||||
Last Closed: | 2019-04-04 11:20:31 UTC | Type: | Bug | ||||||||
Regression: | --- | Mount Type: | --- | ||||||||
Documentation: | --- | CRM: | |||||||||
Verified Versions: | Category: | --- | |||||||||
oVirt Team: | --- | RHEL 7.3 requirements from Atomic Host: | |||||||||
Cloudforms Team: | --- | Target Upstream Version: | |||||||||
Embargoed: | |||||||||||
Bug Depends On: | 1583018 | ||||||||||
Bug Blocks: | |||||||||||
Attachments: |
|
Description
perplexed767
2018-12-17 21:36:55 UTC
Hi, Please share- 1. gluster log 2. geo-replication log from master and 3. mount log from slave. - Sunny (In reply to Sunny Kumar from comment #1) > Hi, > > Please share- 1. gluster log 2. geo-replication log from master and 3. mount > log from slave. > > - Sunny How do I attach logs, or do you wish I email them to you? Created attachment 1516584 [details]
logs-master
logs from master, geo master, brick glustredd, mount logs and changelogs from brick.
Created attachment 1516585 [details]
logs-slave
geo slave logs, brick logs, mount logs, glusterd logs
Created attachment 1516586 [details]
additional info for test volume and geo session.
new test volume and replication session to create requested logs.
Same server, same brick type, lvm xfs, save servers as in description.
Never mind found how to attach logs. I created a new volume and replication session to reproduce the logs you requested. As the other sessions are not created at the moment as they are turned off to do other tests. However same result and very similar setup, except debug logs are turned on and smaller bricks. (In reply to Sunny Kumar from comment #1) > Hi, > > Please share- 1. gluster log 2. geo-replication log from master and 3. mount > log from slave. > > - Sunny Is there any news on this bug? It has been over a month since I filed it. Hi, Looks like https://review.gluster.org/#/c/glusterfs/+/20093/. But I am trying for reproducer to analyse more. If something is missing in step to reproduce please add. - Sunny (In reply to Sunny Kumar from comment #8) > Hi, > > Looks like https://review.gluster.org/#/c/glusterfs/+/20093/. > > But I am trying for reproducer to analyse more. > > If something is missing in step to reproduce please add. > > - Sunny sounds like it could be, which version of gluster is this released in? It is quite easy to reproduce, so don't think I missed anything. default config, replica 2 on both sites, create geo session, start, create and rename a file. You can try this simple test to reproduce the problem. On Master [svc_sp_st_script@hplispnfs30079 conf]$ touch test.txt [svc_sp_st_script@hplispnfs30079 conf]$ vi test.txt a b c d [svc_sp_st_script@hplispnfs30079 conf]$ ll test.txt -rw-r----- 1 svc_sp_st_script domain users 8 Apr 2 14:59 test.txt On Slave [root@hplispnfs40079 conf]# ll test.txt -rw-r----- 1 svc_sp_st_script domain users 8 Apr 2 14:59 test.txt [root@hplispnfs40079 conf]# cat test.txt a b c d On Master [svc_sp_st_script@hplispnfs30079 conf]$ mv test.txt test-moved.txt [svc_sp_st_script@hplispnfs30079 conf]$ ll test-moved.txt -rw-r----- 1 svc_sp_st_script domain users 8 Apr 2 14:59 test-moved.txt On Slave File is not deleted, test-moved.txt does not exist and is not replicated. [root@hplispnfs40079 conf]# ll testfile -rw-r----- 1 svc_sp_st_script domain users 6 Apr 2 14:52 testfile I also tried setting use_tarssh:true but this did not change the behavior. [root@hplispnfs30079 conf]# gluster volume geo-replication common hplispnfs40079::common config access_mount:false allow_network: change_detector:changelog change_interval:5 changelog_archive_format:%Y%m changelog_batch_size:727040 changelog_log_file:/var/log/glusterfs/geo-replication/common_hplispnfs40079_common/changes-${local_id}.log changelog_log_level:INFO checkpoint:0 chnagelog_archive_format:%Y%m cli_log_file:/var/log/glusterfs/geo-replication/cli.log cli_log_level:INFO connection_timeout:60 georep_session_working_dir:/var/lib/glusterd/geo-replication/common_hplispnfs40079_common/ gluster_cli_options: gluster_command:gluster gluster_command_dir:/usr/sbin gluster_log_file:/var/log/glusterfs/geo-replication/common_hplispnfs40079_common/mnt-${local_id}.log gluster_log_level:INFO 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/common_hplispnfs40079_common/gsyncd.log log_level:INFO log_rsync_performance:false master_disperse_count:1 master_replica_count:1 max_rsync_retries:10 meta_volume_mnt:/var/run/gluster/shared_storage pid_file:/var/run/gluster/gsyncd-common-hplispnfs40079-common.pid remote_gsyncd:/usr/libexec/glusterfs/gsyncd replica_failover_interval:1 rsync_command:rsync rsync_opt_existing:true rsync_opt_ignore_missing_args:true rsync_options: rsync_ssh_options: slave_access_mount:false slave_gluster_command_dir:/usr/sbin slave_gluster_log_file:/var/log/glusterfs/geo-replication-slaves/common_hplispnfs40079_common/mnt-${master_node}-${master_brick_id}.log slave_gluster_log_file_mbr:/var/log/glusterfs/geo-replication-slaves/common_hplispnfs40079_common/mnt-mbr-${master_node}-${master_brick_id}.log slave_gluster_log_level:INFO slave_gluster_params:aux-gfid-mount acl slave_log_file:/var/log/glusterfs/geo-replication-slaves/common_hplispnfs40079_common/gsyncd.log slave_log_level:INFO slave_timeout:120 special_sync_mode: ssh_command:ssh ssh_options:-oPasswordAuthentication=no -oStrictHostKeyChecking=no -i /var/lib/glusterd/geo-replication/secret.pem ssh_options_tar:-oPasswordAuthentication=no -oStrictHostKeyChecking=no -i /var/lib/glusterd/geo-replication/tar_ssh.pem ssh_port:22 state_file:/var/lib/glusterd/geo-replication/common_hplispnfs40079_common/monitor.status state_socket_unencoded: stime_xattr_prefix:trusted.glusterfs.bb691a2e-801c-435b-a905-11ad249d43a7.ab3b208f-8cd1-4a2d-bf56-4a98434605c5 sync_acls:true sync_jobs:3 sync_xattrs:true tar_command:tar use_meta_volume:true use_rsync_xattrs:false use_tarssh:true working_dir:/var/lib/misc/gluster/gsyncd/common_hplispnfs40079_common/ This issue is fixed in upstream and 5.x and 6.x series Patch: https://review.gluster.org/#/c/glusterfs/+/20093/ Workaround: The issue affects only single distribute volumes i.e 1*2 and 1*3 volumes. It doesn't affect n*2 or n*3 volumes where n>1. So one way to fix is to convert single distribute to two distribute volume or upgrade to later versions if it can't be waited until next 4.1.x release. REVIEW: https://review.gluster.org/22476 (cluster/dht: Fix rename journal in changelog) posted (#1) for review on release-4.1 by Kotresh HR (In reply to Kotresh HR from comment #13) > This issue is fixed in upstream and 5.x and 6.x series > > Patch: https://review.gluster.org/#/c/glusterfs/+/20093/ We are having the issue in replicate mode (using replica 2). Adrian Sender (In reply to Kotresh HR from comment #14) > Workaround: > The issue affects only single distribute volumes i.e 1*2 and 1*3 volumes. > It doesn't affect n*2 or n*3 volumes where n>1. So one way to fix is to > convert > single distribute to two distribute volume or upgrade to later versions > if it can't be waited until next 4.1.x release. greate thanks, is it planned to be backported to for 4.x as my os (sles 12.2) does not currenty support 5.x gluster) I would have to upgrade the os to sles 12.3 I have backported the patch https://review.gluster.org/#/c/glusterfs/+/22476/. It's not merged yet. REVIEW: https://review.gluster.org/22476 (cluster/dht: Fix rename journal in changelog) merged (#1) on release-4.1 by Kotresh HR I was curious if anyone ever got this resolved? I was running 4.1.7 and set up a geo-replica, this had the above issue with the renaming of files and directories. I have tired up grading to 4.1.8 and have now moved to 5.6 and best I have now is replicated renames of directories. Renaming of files still doesn't get replicated to the geo-replica volume. Some custom compiled RPMS are versioned 4.1.8-0.1.git... and contain the fixes. What a mess this project has become. Broken all versions. Official it will be in 4.1.9 but you can use the below RPMS - we are running it again in production now and appears to be ok. Make sure you update clients as well. [1] RPMs for 4.1 including the fix for el7: https://build.gluster.org/job/rpm-el7/3599/artifact/ -Adrian Sender Thank you Adrian, appreciate the feedback. Unfortunatley that URL returns me a 404 error so cannot get that installed. I may just wait for the 4.1.9 release to go GA, seems based on the release cycle that it mught well be out next week. Unless you happen to have a copy of the RPM's that you can share? Thanks, Ben. This bug is getting closed because a release has been made available that should address the reported issue. In case the problem is still not fixed with glusterfs-4.1.9, please open a new bug report. glusterfs-4.1.9 has been announced on the Gluster mailinglists [1], packages for several distributions should become available in the near future. Keep an eye on the Gluster Users mailinglist [2] and the update infrastructure for your distribution. [1] https://lists.gluster.org/pipermail/gluster-users/2019-June/036679.html [2] https://www.gluster.org/pipermail/gluster-users/ Thx seems to fixed in 6.6 |