2010-03-16 08:39 AM
Yes,that is possible.When the mirror relation is broken between the
two,the secondary becomes writable.Any read/write on that volume is
possible.Issue snapmirror break command from the secondary.Later if the
primary is up and you want to continue the original relation,you can
initiate snapmirror resync to restore the relation.
If you want to permanently break this relation,after snapmirror break,
snapmirror status -l and note the base snapshot name.
Delete the snapshot.You can use snap delete
Edit /etc/snapmirror.conf and delete the entry that defines this
Hope this helps.
2010-03-16 08:45 PM
Thanks a lot ... its was a nice way you solved. One more challange subhada... Why exactly qtress been used? I can create volumes and create share drives. What the needs of Qtress ? Thanks.
2010-03-16 10:01 PM
Qtree is logically defined filesystem within a volume.with quotas
applied on qtrees,you can manage storage resources on a per
Wrt snapmirror,instead of mirroring the entire volume,only the data you
are interested can be mirrored with QSM.Also with snapvault it serves as
a backup copy.
2010-03-17 01:48 AM
Thanks for your answer Shubhada, It is possible to assign a Sharefoler or qtree to a Vfiler. Let me explain my setup here. I have 2 Volumes(Vol0 and Vol1) Vol1 is 10GB, I can install VFiler in that. How can create a sharefolder in Vifiler? Or is it possible to assign a existing sharefolders to Vfilers? Please clarify this.
Thanks in advance...
2010-03-17 04:00 AM
A vfiler is assigned resources (can be either Qtrees or flexvols) which can then have shares / exports made on them as per the normal filer (which is vfiler0 essentially). So to answer the question you can assign an entire volume to the vfiler and create qtrees on it which will all be a used by the vfiler, or you can assign just the qtrees you want to the vfiler and not the volume. Personally, I think the requirement for qtrees is less important now with the advent of flexvols. Thats not to say there is no use case, but flexvols offer better efficiencies with thin provisioning and dedupe etc.
Hope this helps.