Originally posted by: Brentium4
Recently I removed some old LUNs from a collection of Oracle and SQL servers that had been upgraded to new larger LUNs. I had kept the old LUNs around for a week or so to ensure that the DBAs were good with the upgraded disk and didn't want to rollback. I performed the removal task in the AM (between 9-10am). All seemed well in the world, the servers kept chugging along so I didn't reboot them.
At the next monitoring interval I noticed that DLM was complaining of a missing path (makes sense since I removed the disk) but otherwise the servers continued to function normally.
At 6pm backups kicked off for some of the servers and the backup client complained that there was an OS error. Later all all servers that I removed the disks from started having issues with anti-virus. The Network Assosciates Mcshield service started terminating abnormally in a loop. The service would start, then terminate, then because it is set to restart it could continue to restart and fail. In addition to this one of the SQL servers became completely unresponsive.
With so many things going on it would be easy to chalk this up as coincidence. However only the servers (seven of them) that I removed LUNs from were affected. Since no drive letters were assigned to the LUNs that I removed and the server had been rebooted since I unmounted the drives, I have a few questions:
1) Can anyone think of why this would occur?
2) What are your best practices when removing an LDEV/LUN from a server?
Thank you in advance.