[OpenIndiana-discuss] Problems with time-sliderd in b148

[Florian Manschwetus]

I added some datasets (data/backup and all beyond) to an existing pool
(data)(which uses the auto-snapshots since a long time back to b134).
First I disabled the snapshotting of this dataset by setting
com.sun:auto-snapshot locally to false. Now after some testing I set it
locally to true (later I tried to inherit but no difference).
When snapshotting data/backup is enabled it leads to a crash / restart
of time-sliderd as he tries to snapshot the leave-datasets twice for one
scheduled event. That ends up it just having snapshots from the farest
leaves and nothing else (even the all the time working hourlys of
data/Admin aren't made then). When setting backup back to false the rest
works fine again.

Any idea out there?

Thx,
Florian


Here a short extract from the service log:

--- snip ---

Failed to create snapshots for schedule: hourly
Caught RuntimeError exception in snapshot manager thread
Error details:
--------BEGIN ERROR MESSAGE--------
['/usr/bin/pfexec', '/usr/sbin/zfs', 'snapshot', '-r',
'data/backup/serviceDaten at zfs-auto-snap_hourly-2011-05-05-09h13'] failed
with exit code 1
cannot create snapshot
'data/backup/serviceDaten/windows at zfs-auto-snap_hourly-2011-05-05-09h13': dataset
already exists
no snapshots were created

--------END ERROR MESSAGE--------
Snapshot monitor thread exited.
[ Mai  5 09:13:06 Stopping because all processes in service exited. ]

--- snip ---