9:01 p.m.
于 2013-1-29 4:00, Dan Kenigsberg 写道:
Following are the notes that I've taking during the call.
I'm pretty sure that I'm missing something important, such as another blocker
BZ, so please reply with corrections/addtions.
- Adam: going to introduce Python binding for the new Vdsm API
- Sharad has joined our call for the first time. Welcome!
Interested in backup/restore functionality in oVirt, to facilitate
integration with TSM, presumably
http://www-142.ibm.com/software/products/us/en/tivostormana/
- Federico warns that live snapshot is not enough, since we lack live
merge. After a year of daily backups, each qcow chain would be 365
hops deep.
That's horrible. It looks like that Qemu 1.3 and libvirt 1.0.0 already
supports the "live block commit" which is the base of live commit. Do
we have plan to support "live commit" in next oVirt release?
ovirt-3.2:
- We should probably revert http://gerrit.ovirt.org/9315 (integrate
zombie reaper in supervdsmServer) as it tickles a bug in
multiprocessing. See
http://lists.ovirt.org/pipermail/users/2013-January/011857.html for
more details in the thread starting in
http://lists.ovirt.org/pipermail/users/2013-January/011747.html
(thanks for raising the issue and finding the culprit, Dead Horse)
- Danken believes that all important el6-specific issues reported by dreyou are
now handled in the master and ovirt-3.2 branches. Please report this list if
he's wrong.
- Federico reports that udev is going to revert to its Fedora 18alpha
behavior so that we can keep our integration with it. I'd love to see
the udev BZ listed in our tracker
https://bugzilla.redhat.com/showdependencytree.cgi?id=881006&hide_res...
as I do not recall it. Federico, could you add it? We should probably
require a newer udev release.
- Federico reports that we must take a short-but-important patch from Lee
Yarwood: http://gerrit.ovirt.org/#/c/11281/
Cross-distro:
- Adam: a guy from IBM is now listing issues that block running Vdsm on
Ubuntu. Yay!
- Saggi suggests that IBM contribute an Ubuntu slave for Jenkins, so
that each and every patch is tested not to introduce Ubuntu
regressions - just as Red Hat has recently done with EL6.
- Danken: yes, we have got el6-based unit-testing running. If you are still
getting an unjustified X, you may need to rebase on current master.
- Federico: NetworkManager 9 has bridge support, which leads the way to
NM-based implementation of Vdsm's configNetwork.
- Danken: Toni is working on a suggestion for refactoring configNetwork
in such a way that multiple implementations (e.g. ifcfg-based,
NM-based) can coexist. Stay stuned for his Wiki page on the subject.
Other refactoring going on:
- Vinzenz has begun breaking the horrible libvirtvm+vm to edible-size
pieces. Reviews are most welcome, particularly from Mark Wu, whose
http://gerrit.ovirt.org/10054/ only awaits verification tick.
Functional Testing:
- Adam: a guy from IBM (Zhou, I presume) is running the functional tests
via Jenkins on his laptop. Yay!
Zhou Zheng Sheng <zhshzhou(a)linux.vnet.ibm.com>
- Adam: plans to add a functional test for the getVdsCaps verb, as
he
found out that an evil Vdsm developer has added values to the caps
without updating the schema.
- P.S. FlowID is now in for storage verbs as well (thanks, Douglas). I
hope Haim is happier now.
Happy coding!
_______________________________________________
vdsm-devel mailing list
vdsm-devel(a)lists.fedorahosted.org
https://lists.fedorahosted.org/mailman/listinfo/vdsm-devel
--
---
舒明 Shu Ming
Open Virtualization Engineerning; CSTL, IBM Corp.
Tel: 86-10-82451626 Tieline: 9051626 E-mail: shuming(a)cn.ibm.com or
shuming(a)linux.vnet.ibm.com
Address: 3/F Ring Building, ZhongGuanCun Software Park, Haidian District, Beijing 100193,
PRC