I did the upgrade on two servers. Both of them brought the pools online, but one of them picked /dev/sdXX, the other picked by-id/wwn-*.
They are identical machines (DL380g9 both with 12x6tb SAS Drives), so the discovered devices are effectively random. I assume /dev/disk/by- id/scsi-3500* would also be in the pool of things it could have chosen. However, once the devices are discovered 'missing' and zfs tries to find them, they are then effectively stuck as that until the pool is exported and re-imported. As Doug mentioned, /dev/sdxx is strongly discouraged because of the randomness of device discovery, and also significantly increases boot delay if every disk needs to be probed. This can take many minutes on a machine with a large number of spindles - I have a bulk storage server with 169 disks (2 x 84 bay JBODs) that has taken 7 minutes to just load its zpools during boot when an equivalent thing happened to it. -- You received this bug notification because you are a member of Ubuntu Bugs, which is subscribed to Ubuntu. https://bugs.launchpad.net/bugs/2137191 Title: upstream 00-scsi-sg3_config.rules missing from distro package To manage notifications about this bug go to: https://bugs.launchpad.net/ubuntu/+source/sg3-utils/+bug/2137191/+subscriptions -- ubuntu-bugs mailing list [email protected] https://lists.ubuntu.com/mailman/listinfo/ubuntu-bugs
