Hi,
2 days ago I upgraded a server that is used as backup machine to 8.2 prerelease. All seemed to be ok after the upgrade, [cmd=]zpool upgrade -a[/cmd] was done immediately and [cmd=]zfs upgrade -a[/cmd] has taken some time, a hour and some minutes.
After this I started a scrub with the knowledge that it takes hours to be finished on this pool and I knew it would hit the time window, when some backups are stored to that pool, using rsync over the network. I decided that this would became a good stress test, because it worked several times and generates heavy load, without problems on FreeBSD 8.1 AMD64.
The server crashed with a double fault. Unfortunately I cannot provide more information about this and it was not a GENERIC kernel. Next time the server will crash, I will collect everything necessary.
That's the story, now my questions:
Did anyone experience anything similar after upgrading?
Is it foolproof that [cmd=]zfs upgrade -a[/cmd] works on an "online" pool with maybe some thousands of mountpoints?
Greetings & thanks in advance
2 days ago I upgraded a server that is used as backup machine to 8.2 prerelease. All seemed to be ok after the upgrade, [cmd=]zpool upgrade -a[/cmd] was done immediately and [cmd=]zfs upgrade -a[/cmd] has taken some time, a hour and some minutes.
After this I started a scrub with the knowledge that it takes hours to be finished on this pool and I knew it would hit the time window, when some backups are stored to that pool, using rsync over the network. I decided that this would became a good stress test, because it worked several times and generates heavy load, without problems on FreeBSD 8.1 AMD64.
Code:
NAME SIZE USED AVAIL CAP HEALTH ALTROOT
backup1 16.3T 7.19T 9.13T 44% ONLINE -The server crashed with a double fault. Unfortunately I cannot provide more information about this and it was not a GENERIC kernel. Next time the server will crash, I will collect everything necessary.
That's the story, now my questions:
Did anyone experience anything similar after upgrading?
Is it foolproof that [cmd=]zfs upgrade -a[/cmd] works on an "online" pool with maybe some thousands of mountpoints?
Greetings & thanks in advance
