This is intended to track progress on both servers, but each will have its own task since they will need to be done rather carefully, one at a time, in order to ensure minimal interruption of services. labstore1004/5 (the primary NFS cluster) both have unused dual port 10G interfaces (T127508). This is for the primary traffic interface first, and then the crossover cable.
labstore1004 is in rack C2, which I believe is a 10G rack, though it seems fairly busy. If switch ports are available, that makes that server somewhat simpler, though the crossover cable would need to be moved to the second 10G port on the server as well.
labstore1005 is in rack C5, which means it needs to be moved to a different rack altogether in order to set it up as 10G.
labstore1005 requires stopping replication from 1004 and stopping the backup jobs from running before shutdown.
labstore1004 requires a failover to labstore1005 to proceed.
I believe that the procedure used to re-image cloudvirt1019 and 1020 could be used in order to preserve the data volumes.
- labstore1005 T266199
- labstore1004 T266202
- Reconnect the DRBD "crossover" cable setup between both servers with 10G interfaces to bring write performance to full levels.