Thanks for your response!
I have the following two questions now, please help me
1) There is one
node for every cluster. Ceph and radosgw are deployed at every node.
One is as master zone, another is as slave zone. I write data to master
zone with bash script, and run radosgw agent at slave zone. I use the bash script composed of “s3cmd” to write data to the master
zone with block of 4k.
I found that the speed of writing data to the
master zone is faster than syncing data to slave zone for incremental sync. I have increase the number of worker and
decrease the incremental-sync-delay time, but it is of no effect. Could you have other mode to increase the speed
of syncing data to the slave zone?
2) If I reverse the
role for the master zone and slave zone, radosgw agent
will copy data from the new master zone. The data to sync is actually exist in the
new slave zone already. Because there is no replica_log in the new slave zone, radosgw agent will sync data repeatedly. Could you have any method to avoid this?
Thanks !
Fangchen
2015-11-28 12:45 GMT+08:00 Yehuda Sadeh-Weinraub <ysadehwe@xxxxxxxxxx>:
In theory yes, but I wouldn't recommend that as it's severely under tested. Good chance that the agents will step on each other's toes. You can modify there number of concurrent threads for the agents, might be enough for your requirements.
Yehuda
On Nov 27, 2015 2:40 AM, "fangchen sun" <sunfangchen2008@xxxxxxxxx> wrote:_______________________________________________HI All,I have two ceph cluster. They are configured with federation. I use radosgw-agent to sync data between the two cluster. Can I deploy more than one radosgw-agent to sync data simultaneously?Thanks！
ceph-users mailing list
ceph-users@xxxxxxxxxxxxxx
http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com
_______________________________________________ ceph-users mailing list ceph-users@xxxxxxxxxxxxxx http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com
