I have noticed for a while that "zfs promote <clone>" doesn't seem to
complete properly, even though no error is returned.
I'm still running the nevada builds and use lucreate/luupgrade/ludelete
to manage my BE:s. Due to issues with ludelete failing to clean up old
BE:s properly I've made it a habit of promoting the oldest BE I want to
keep before I run ludelete, and more often than not do I have to run the
promote twice before it is complete, and both runs return good status.
Here is a run I just did where I wanted to get rid of snv_125 and keep
snv_126 (and later):
---
spiff:~ (334)> zfs promote rpool/ROOT/snv_126
spiff:~ (335)> zfs promote rpool/ROOT/snv_126
spiff:~ (336)> zfs promote rpool/ROOT/snv_126
cannot promote 'rpool/ROOT/snv_126': not a cloned filesystem
---
Notice that the first promote appears to have run OK, but the next also
runs fine, and not until I try it a third time do I get an error which
indicates that this is no longer a clone.
I assume the problem is related to auto-snapshot, which I run.
Someone suggested that the problem might be that I run the promote while
the system is busy taking a snapshot, but the run above was done 5
minutes before the next snapshot should have been run, and 10 minutes
after the last snapshot, so this should not have been the problem.
Is this a known problem?
Thomas
_______________________________________________
zfs-discuss mailing list
zfs-discuss@opensolaris.org
http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
