| mcape | Hello all, my name is Oleksandr and I'm from Ukraine. I work for ukrainian government procurement system, and all documents are stored in openstack swift storage, about 120M of files. | 11:48 |
|---|---|---|
| mcape | We did tried to rescue the data, and to do that, we disconnected one zone (holding one of three replicas) and moved it to city on the west of the country, to a safer place. | 11:48 |
| mcape | We removed that zone from the ring as well and changed replica number to two. There was about 1% of partition movement on that rebalance. | 11:48 |
| mcape | Now, the servers are safely arrived to Lviv. | 11:48 |
| mcape | I'm asking you to validate my thinking - my plan is just restore old three replicas configuration (we have backup) and replication will fix everything. Is it correct? | 11:48 |
| timburke__ | mcape, yes, i'd expect restoring the backup ring should allow replication to bring everything back up to fully (triple replica) durability quickly | 22:51 |
| timburke__ | note that you'll want to double-check that the IP addresses of all the nodes in the moved zone are the same (or update them after rolling back the builder but before deploying the restored ring) | 22:51 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!