Both an older Dell ML6000 and a newer HP MSL3040 keep losing robotics communication with NetBackup. Hardware diagnostics haven't found a problem, but rebooting the libraries keeps being necessary. Where should I look next?
Why do my tape libraries keep needing reboots for robotics errors?
Answers
Two different libraries showing the same symptom would make me look at the shared host and backup software side first. That doesn't prove the libraries are fine, but it gives you a useful place to start.
Check the library and drive firmware, HBA firmware and drivers against the compatibility information for your NetBackup version. Also check whether the changer has a stable device path or persistent binding, especially if the trouble follows a host or SAN event.
Reply to this answer
How are they connected, SAS or Fibre Channel? If they're on a SAN, check zoning and whether another host is probing or claiming the changer LUN. A reservation conflict could leave the robotics stuck, so it's worth checking what else can see the device.
Reply to this answer
Before power cycling the library next time, see whether restarting the affected server restores communication. That may help narrow down which side is stuck, although it won't by itself identify the cause.
Reply to this answer
Frequent reboots aren't something I'd just accept as normal operation. Get the library vendor and backup software support involved, particularly if the host and configuration checks don't turn up anything. Having both models affected is useful context for the support case.
Reply to this answer
Add your answer
Answers are reviewed before publication.
8m+fablemortar_527:
That could also explain why a library reboot clears the symptom without fixing what caused it.