Blitz: Setup vSphere Replication (SRM 5.0)
- Do ONE side at a time.
- Each side must be treated as if it was separate – linked mode does not work for setting up VRMS. That means a VI client to Side A while setting up Side A, and a client open to Side B while setting up Side B. This is because the certificate warnings do not propagate through the link between VCs .
- Don’t use DNS in VRMS. Use IP addresses only for everything in VRMS config, regardless of how well DNS works in the environment.
- Static IPs cannot be changed after being set. This is a known PR.
- Static IPs work better than DHCP for VRMS servers, if at all possible.
- DB2 works better than SQL.
- SQL connections may not be closed on reboot, requiring a restart of the DB. You may want have a separate SQL server for VRMS because of this.
- Old Certs don’t work.
- Any VC2.5 originating / VC4.0 cert uses MD5onRSA, which is incompatible with VRMS.
- It will accept the cert anyway, but will not connect from experience.
- Either build a new VC, or update to a new 2048 bit certificate. Working on details of requirements now
- Setup order is VERY specific.
- Deploy Site “A” VRMS server from VI Client to Site “A”.
- Open web browser to IP of VRMS server “A”.
- Generate new SSL Certificate and install.
- Add settings for VC / DB / etc, BY IP ONLY.
- Hit “Save and Restart Service” when setup.
- On Site “A” VI Client, open “vCenter Solutions Manager”
i. Click VR Management. You should be soon prompted to accept a certificate. Do so, and click the ignore button.
- Open SRM. Click vSphere Replication for the appropriate site.
i. Wait for a certificate prompt. Accept it.
ii. In about a minute, you should see the VRMS server log in to SRM.
- Close VI Client for site “A”, open VI Client to site “B”, and repeat prior steps on Site “B”.
- Connect VRMS servers together.
- Deploy VR servers on required systems.
- Add VR servers.
- Configuring VMs for VRMS replication.
- If initial configuration fails with a “optimistic locking failure” immediately try to reconfigure again.
- If the second time it fails with a “VC View” error, remove replication and try to configure fresh after ~10 seconds.
- IMMEDIATELY after configuring replication, go to SRM to verify that it has started the “initial sync”.
- It seems that if this step is skipped, VR may decide that the VM is synced, even though it isn’t (reproduced twice, verifying – it claims to have replicated a 15g vm in 10 seconds).
- If for some reason it does that, remove replication, reconfigure replication, and set RPO to 15 minutes. This seems to always trigger the initial sync correctly. Modify RPO as needed afterwards.
- SRM hangs – linked mode tends to cause VC to close connections. Check number of logged in sessions if this happens.