Niall Power wrote: > Bill Shannon wrote: >> Niall Power wrote: >>> Bill Shannon wrote: >>>> Niall Power wrote: >>>>> The zfssnap role entry in /etc/user_attr should look like this: >>>>> >>>>> zfssnap::::type=role;auths=solaris.smf.manage.zfs-auto-snapshot;profiles=ZFS >>>>> >>>>> File System Management >>>> >>>> If I add that entry, how do I restart/reinitialize everything? >>>> Can I just "svcadm restart time-slider"? >>> Run "svcs -xv" to see if any of the services are in maintenance mode >>> such as "auto-snapshot:hourly" and then clear them by running >>> "svcadm clear hourly", "svcadm clear frequent" , etc. >>> This should bring everything back online. >> >> They're all already online. Should I restart them all? > Then they should be working fine and no need to take further action. > If: > svc:/application/time-slider:default > reports as being online also then you're in business.
I double checked and I guess I was wrong. Some are online and some aren't: $ svcs | egrep 'zfs|time' online Dec_05 svc:/system/filesystem/zfssnap-roleadd:default online Dec_05 svc:/network/shares/group:zfs online Dec_10 svc:/system/filesystem/zfs/auto-snapshot:weekly online Dec_10 svc:/system/filesystem/zfs/auto-snapshot:monthly offline Jan_06 svc:/application/time-slider:default maintenance Jan_06 svc:/system/filesystem/zfs/auto-snapshot:frequent maintenance Jan_06 svc:/system/filesystem/zfs/auto-snapshot:hourly maintenance 0:00:04 svc:/system/filesystem/zfs/auto-snapshot:daily I did "svcadm clear" on the ones in maintenance mode, but it didn't work. The log files show: cannot create snapshot 'rpool/ROOT/opensolaris at zfs-auto-snap:daily-2009-01-07-15 :13': dataset is busy Now what?
