Hi,
Thanks for reporting the issue. This seems to be a bug.On Wed, Feb 21, 2018 at 2:01 PM, Marcus Pedersén <marcus.pedersen@xxxxxx> wrote:
Hi all,
I use gluster 3.12 on centos 7.
I am writing a snapshot program for my geo-replicated cluster.
Now when I started to run tests with my application I have found
a very strange behavior regarding geo-replication in gluster.
I have setup my geo-replication according to the docs:
http://docs.gluster.org/en/latest/Administrator%20Guide/ Geo%20Replication/
Both master and slave clusters are replicated with just two
machines (VM) and no arbiter.
I have setup a geo-user (called geouser) and do not use
root as the geo user, as specified in the docs.
Both my master and slave volumes are named: vol
If I pause the geo-replication with:
gluster volume geo-replication vol geouser@ggluster1-geo::vol pause
Pausing geo-replication session between vol & geouser@ggluster1-geo::vol has been successful
Create a snapshot:
gluster snapshot create my_snap_no_1000 vol
snapshot create: success: Snap my_snap_no_1000-2018.02.21-07.45.32 created successfully
Resume geo-replication:
gluster volume geo-replication vol geouser@ggluster1-geo::vol resume
Resuming geo-replication session between vol & geouser@ggluster1-geo::vol has been successful
Everything works fine!
But here comes the problem:
If I by accident spell my slave user wrong or don't use
the user at all, as I was using root,
no matter what user I write pause/resume do NOT report
any errors. The answer is always pausing/resuming successful.
The problem comes after a successful pause when I try to
create a snapshot. It fails with:
snapshot create: failed: geo-replication session is running for the volume vol. Session needs to be stopped before taking a snapshot.
gluster volume geo-replication status
MASTER NODE MASTER VOL MASTER BRICK SLAVE USER SLAVE SLAVE NODE STATUS CRAWL STATUS LAST_SYNCED
------------------------------------------------------------ ------------------------------ ------------------------------ -------------------------
ggluster1 vol /gluster geouser ssh://geouser@ggluster1-geo::vol N/A Paused N/A N/A
ggluster2 vol /gluster geouser ssh://geouser@ggluster1-geo::vol N/A Paused N/A N/A
After this snapshots fails all the time!
If I use the correct user again and pause, no error (paused), snapshot fails.
If I resume with correct user, no errors (active).
Geo-replication still works fine, but some how has something
gone wrong so snapshots fail.
After restart of glusterd in all machines it starts to work again.
Here is complete run through:
gluster volume geo-replication status
MASTER NODE MASTER VOL MASTER BRICK SLAVE USER SLAVE SLAVE NODE STATUS CRAWL STATUS LAST_SYNCED
------------------------------------------------------------ ------------------------------ ------------------------------ ------------------------------ ----------
ggluster1 vol /gluster geouser ssh://geouser@ggluster1-geo::vol ggluster1-geo Active Changelog Crawl 2018-02-12 15:49:57
ggluster2 vol /gluster geouser ssh://geouser@ggluster1-geo::vol ggluster2-geo Passive N/A N/A
# Using wrong user: abc
gluster volume geo-replication vol abc@ggluster1-geo::vol pause
Pausing geo-replication session between vol & abc@ggluster1-geo::vol has been successful
gluster volume geo-replication status
MASTER NODE MASTER VOL MASTER BRICK SLAVE USER SLAVE SLAVE NODE STATUS CRAWL STATUS LAST_SYNCED
------------------------------------------------------------ ------------------------------ ------------------------------ -------------------------
ggluster1 vol /gluster geouser ssh://geouser@ggluster1-geo::vol N/A Paused N/A N/A
ggluster2 vol /gluster geouser ssh://geouser@ggluster1-geo::vol N/A Paused N/A N/A
gluster snapshot create snap_vol_1000 vol
snapshot create: failed: geo-replication session is running for the volume vol. Session needs to be stopped before taking a snapshot.
# Using wrong user: abc
gluster volume geo-replication vol abc@ggluster1-geo::vol resume
Resuming geo-replication session between vol & ggluster1-geo::vol has been successful
gluster volume geo-replication status
MASTER NODE MASTER VOL MASTER BRICK SLAVE USER SLAVE SLAVE NODE STATUS CRAWL STATUS LAST_SYNCED
------------------------------------------------------------ ------------------------------ ------------------------------ ------------------------------ ----------
ggluster1 vol /gluster geouser ssh://geouser@ggluster1-geo::vol ggluster1-geo Active Changelog Crawl 2018-02-12 15:49:57
ggluster2 vol /gluster geouser ssh://geouser@ggluster1-geo::vol ggluster2-geo Passive N/A N/A
Many thanks in advance!
Best regards
Marcus
--
**************************************************
* Marcus Pedersén *
* System administrator *
**************************************************
* Interbull Centre *
* ================ *
* Department of Animal Breeding & Genetics — SLU *
* Box 7023, SE-750 07 *
* Uppsala, Sweden *
**************************************************
* Visiting address: *
* Room 55614, Ulls väg 26, Ultuna *
* Uppsala *
* Sweden *
* *
* Tel: +46-(0)18-67 1962 *
* *
**************************************************
* ISO 9001 Bureau Veritas No SE004561-1 *
**************************************************
_______________________________________________
Gluster-users mailing list
Gluster-users@xxxxxxxxxxx
http://lists.gluster.org/mailman/listinfo/gluster-users
--
Thanks and Regards,
Kotresh H R_______________________________________________ Gluster-users mailing list Gluster-users@xxxxxxxxxxx http://lists.gluster.org/mailman/listinfo/gluster-users
