[OmniOS-discuss] LTS just comes to a halt doing mass-copy

Dan McDonald danmcd at omniti.com
Mon Apr 27 04:55:24 UTC 2015


> On Apr 25, 2015, at 9:14 PM, Michael Mounteney <gate03 at landcroft.co.uk> wrote:
> 
> Supplementary:
> 
> 1.  To revert, I booted from c2t2d0s0 then zpool replaced c2t0d0s0.
> That went OK.

Okay.

> 2.  The mass copying process (step 5 of TFA) died on one particular zfs
> filesystem, which had child systems, although it had already
> successfully copied another filesystem with child systems.  I restarted
> the machine and resumed mass-copying and it again died on the same
> filesystem.  Is it possible that the system contains features that LTS
> can't handle ?  Surely LTS would say so ?

By "died" you mean the system hung?  Did you have a concurrent shell running at the same time?  Did that stop working?  If you have a concurrent shell, it might be nice to take a system dump upon this freeze.

> 3.  Having reverted and replaced, I then copied the offending
> filesystem within the same rpool, i.e.:
> 
> root at world:/root# zfs snapshot -r rpool/gentoo at Z
> root at world:/root# nice zfs send -R rpool/gentoo at Z | nice zfs recv rpool/G2
> 
> which went alright, so the filesystem itself does not appear to be
> unreadable.
> 
> Does the above provide any clues ?

Where was the original zpool coming from?  You mention an "upgrade" so I'm assuming it's from an earlier version of OmniOS.

Can you send the snapshot to a file?  It might be useful to debug the snapshot.

Dan



More information about the OmniOS-discuss mailing list