On Tue, May 7, 2024 at 7:54 AM Eugen Block <eblock@nde.ag> wrote:
Hi,
I'm not the biggest rbd-mirror expert. As understand it, if you use one-way mirroring you can failover to the remote site, continue to work there but there's no failover back to primary site. You would need to stop client IO on DR, demote the image and then import the remote images back to primary site. Once everything is good you can promote the image on primary again. The rbd-mirror will then most likely be in a split-brain situation, which can be resolved by resyncing images from primary again. You can't do a resync on primary site because there's no rbd-mirror daemon running.
Having two-way mirroring could help, I believe. Let's say you lose the primary site, you can (force) promote images on the remote site, continue working. Once the primary site is back up (but not primary yet), you can do the image resync from the remote (currently primary) site (because there's a rbd-mirror daemon running on the primary site as well). Once the primary site has all images promoted, you'll probably have to resync on the remote site again to get out of the split-brain.
Also, you need to demote the out-of-date images in the cluster that came back, before issuing resync on them. This is to resolve the split-brain. See, https://docs.ceph.com/en/latest/rbd/rbd-mirroring/#force-image-resync -Ramana
But at least you won't need to export/import images.
But you'll need to test this properly to find out if your requirements are met.
Regards, Eugen
Zitat von V A Prabha <prabhav@cdac.in>:
Dear Eugen, We have a scenario of DC and DR replication, and planned to explore RBD mirroring with both Journaling and Snapshot mechanism. I have a 5 TB storage at Primary DC and 5 TB storage at DR site with 2 different ceph clusters configured.
Please clarify the following queries
1. With One way mirroring, the failover works fine in both journaling and snapshot mechanism and we are able to promote the workload from DR site. How does Failback work? We wanted to move the contents from DR to DC but it fails. In journaling mechanism, it deletes the entire volume and recreates it afresh which does not solve our problem. 2. How does incremental replication work from DR to DC? 3. Does Two-way mirroring help this situation. According to me, in this method, it is for 2 different clouds with 2 different storages and replicating both the clouds workloads? Does Failback work in this scenario ? Please help us / guide us to deploy this solution
Regards V.A.Prabha
Thanks & Regards, Ms V A Prabha / श्रीमती प्रभा वी ए Joint Director / संयुक्त निदेशक Centre for Development of Advanced Computing(C-DAC) / प्रगत संगणन विकास केन्द्र(सी-डैक) Tidel Park”, 8th Floor, “D” Block, (North &South) / “टाइडल पार्क”,8वीं मंजिल, “डी” ब्लॉक, (उत्तर और दक्षिण) No.4, Rajiv Gandhi Salai / नं.4, राजीव गांधी सलाई Taramani / तारामणि Chennai / चेन्नई – 600113 Ph.No.:044-22542226/27 Fax No.: 044-22542294 ------------------------------------------------------------------------------------------------------------ [ C-DAC is on Social-Media too. Kindly follow us at: Facebook: https://www.facebook.com/CDACINDIA & Twitter: @cdacindia ]
This e-mail is for the sole use of the intended recipient(s) and may contain confidential and privileged information. If you are not the intended recipient, please contact the sender by reply e-mail and destroy all copies and the original message. Any unauthorized review, use, disclosure, dissemination, forwarding, printing or copying of this email is strictly prohibited and appropriate legal action will be taken. ------------------------------------------------------------------------------------------------------------
_______________________________________________ ceph-users mailing list -- ceph-users@ceph.io To unsubscribe send an email to ceph-users-leave@ceph.io