Created attachment 1331213 [details]
Description of problem:
In case I log into new targets when adding new domain but not finish the process the new targets will not be logged out when the host will put into maintenance.
In case I do add a new storage domain, all targets that the host was logged into them will be cleared when the host will move to maintenance
Version-Release number of selected component (if applicable):
Steps to Reproduce:
1. Open new domain dialog and login to few targets
2. Cancel the operation
3. Verify using '# iscsiadm -m session' that the new targets were added
4. Put host into maintenance and verify that the new targets from step 3 are still there
5. with new host repeat step 1 and finish the addition of new storage domain
6. put host into maintenance and verify that the target got cleared
This bug report has Keywords: Regression or TestBlocker.
Since no regressions or test blockers are allowed between releases, it is also being identified as a blocker for this release. Please resolve ASAP.
I don't think this is a regression.
It's obviously not the best behavior, to say the least, but I think it's always been like that.
On what version did you see this working as you expect?
(In reply to Allon Mureinik from comment #2)
> I don't think this is a regression.
> It's obviously not the best behavior, to say the least, but I think it's
> always been like that.
> On what version did you see this working as you expect?
(In reply to Raz Tamir from comment #3)
> (In reply to Allon Mureinik from comment #2)
> > I don't think this is a regression.
> > It's obviously not the best behavior, to say the least, but I think it's
> > always been like that.
> > On what version did you see this working as you expect?
Logs of such a rub please?
Seems like my environment was in a bad state when open this bug - This is not a regression.
In 4.1 this works the same.
Also, bug #1496206, which is in the same area is not a regression
The only real solution for this bug is providing a proper way to manage storage server connections, i.e the relationship between hosts, connections and LUNs.
This actually is an RFE and not a bug. It has a workaround (logging out from the unwanted targets) and thus should, IMHO, be closed/deferred. Tal?
The solution here seems too complicated and the impact is quite low, I don't think it worth the effort