Milan Zamazal has posted comments on this change.

Change subject: virt: Initial support for post-copy migration
......................................................................


Patch Set 10:

(4 comments)

https://gerrit.ovirt.org/#/c/62873/8/vdsm.spec.in
File vdsm.spec.in:

Line 187: Requires: libvirt-daemon-kvm >= 1.3.5
Line 188: %else # rhel
Line 189: Requires: libvirt-daemon-kvm >= 1.3.3
Line 190: Requires: libvirt-python >= 1.3.4
Line 191: %endif
> ok, this may need to be split in a separate change, depending on the time o
We'll see. The problem is that once we removed the backward compatibility code 
from this patch, we must require corresponding libvirt versions, otherwise we 
crash on libvirt.VIR_MIGRATE_POSTCOPY (we can make a workaround if it's still a 
problem at the time of merge).
Line 192: 
Line 193: # iscsi-intiator versions
Line 194: %if 0%{?rhel}
Line 195: Requires: iscsi-initiator-utils


https://gerrit.ovirt.org/#/c/62873/6/vdsm/virt/migration.py
File vdsm/virt/migration.py:

Line 480:                      (libvirt.VIR_MIGRATE_COMPRESSED if
Line 481:                          self._compressed else 0) |
Line 482:                      (libvirt.VIR_MIGRATE_AUTO_CONVERGE if
Line 483:                          self._autoConverge else 0))
Line 484: 
> Vdsm *should* be run in F24, or, to word it differently, I'm not aware of a
OK, thanks for confirming we should be fine in this regard.
Line 485:             self._vm._dom.migrateToURI3(duri, params, flags)
Line 486:         else:
Line 487:             self._raiseAbortError()
Line 488: 


https://gerrit.ovirt.org/#/c/62873/8/vdsm/virt/migration.py
File vdsm/virt/migration.py:

Line 478
Line 479
Line 480
Line 481
Line 482
> unrelated, please drop this change if you resubmit
Done


PS8, Line 778: self._vm.log.info('Switching to post-copy mig
> ugh, this reads so C-ish :)
Well, I don't know how to identify the C-ishness of the condition. Thinking 
about it I tend to use Lisp style and to avoid Prolog style (to make clear we 
are in the procedural and not declarative programming area). I admit I do some 
C++ programming, but I may be blamed for excessive use of virtual methods 
there, so I'd say I'm more influenced by Python than C/C++. Dancing among those 
subtle and delicate issues, I may indeed overlook that I use C-ish constructs 
in completely inappropriate places.

So I changed it as you suggest, to be consistent with the places where we use a 
similar construct.


-- 
To view, visit https://gerrit.ovirt.org/62873
To unsubscribe, visit https://gerrit.ovirt.org/settings

Gerrit-MessageType: comment
Gerrit-Change-Id: I4c9911f47331120a1b78326044a4949abf35d5fe
Gerrit-PatchSet: 10
Gerrit-Project: vdsm
Gerrit-Branch: master
Gerrit-Owner: Milan Zamazal <[email protected]>
Gerrit-Reviewer: Francesco Romani <[email protected]>
Gerrit-Reviewer: Jenkins CI
Gerrit-Reviewer: Michal Skrivanek <[email protected]>
Gerrit-Reviewer: Milan Zamazal <[email protected]>
Gerrit-Reviewer: Tomas Jelinek <[email protected]>
Gerrit-Reviewer: gerrit-hooks <[email protected]>
Gerrit-HasComments: Yes
_______________________________________________
vdsm-patches mailing list -- [email protected]
To unsubscribe send an email to [email protected]

Reply via email to