Vdsm sync call April 1st 2014 =============================
- cpopen: - Toni: there's a versioning mismatch: the version in pypi is higher than the one on https://github.com/ficoos/cpopen - Saggi: there shouldn't be any code difference - Yaniv should sync the two.
- We use two options that are missing from Python3's Popen: umask and deathSignal. - Toni to re-send his Python3 port for cpopen, with tests running on Python3, too. - Saggi to accept it. - Infra team to suggest missing features to Python.
- fromani described his attempts of consolidating the two migration-monitoring threads into one. The motivation is a finer way of contolling the migration down time based on progress. Reducing thread numbers per se is not the main motivation.
- pep8 has changed. Again. Version 1.5.1 has even more draconic requirements. We can comply with them, or, include our version of pep8/pyflakes/pylint in our git repo. danken shudders at the thought of copying the code, but having a git sub-module is a reasonable compromise. - Infra team to take this task (unless someone else is faster) - Until that happens, one need to use pep8-1.4.6 or --ignore offending errors.
- We've been asked to kill the separate vdsm-devel mailing list, and join it into devel@ovirt.org. There's some resistence in the audience, so we'd do it softly.
Next week, I'll have vdsm-devel members mass-subscribed to devel@ovirt. If you do NOT want to be subscribed, please send me a private request.
Then, for several months, we'd see how it goes, and whether we drown in unrelated Engine chatter.
- We had a very (too) heated debate about ignoring failures of setDomainRegularRole() in http://gerrit.ovirt.org/24495/ and http://gerrit.ovirt.org/25424.
The pain point is that relying on domain role (master/regular) is faulty by design. We cannot avoid the cases where a pool has more than one domain with a master role written in its metadata.
One side argued that oVirt should be fixed to handle this unescapable truth, or at least enumerate the places where Vdsm and Engine, both current and old, depend on master role uniqueness.
The other side argued that this is not a priority task, and that we should try to "garbage-collect" known-bad master roles as a courtesy to people digging into domain metadata, and as a means to lessen the problem until we kill the pool concept in the upcoming version.
I hope that I present the debate fairly enough.
Dan.
----- Original Message -----
From: danken@redhat.com To: "VDSM Project Development" vdsm-devel@lists.fedorahosted.org, devel@ovirt.org Sent: Wednesday, April 2, 2014 10:29:09 AM Subject: [vdsm] vdsm sync meeting minutes April 1st, 2014
Vdsm sync call April 1st 2014
cpopen:
Toni: there's a versioning mismatch: the version in pypi is higher than the one on https://github.com/ficoos/cpopen
Saggi: there shouldn't be any code difference
Yaniv should sync the two.
We use two options that are missing from Python3's Popen: umask and deathSignal.
- Toni to re-send his Python3 port for cpopen, with tests running on Python3, too.
https://github.com/celebdor/cpopen/commit/6db0429de06020c2c62007f6d1a3b41906...
I did it last night. There's still the noclosefds test failing for unknown reasons. I'll investigate it.
- Saggi to accept it. - Infra team to suggest missing features to Python.
fromani described his attempts of consolidating the two migration-monitoring threads into one. The motivation is a finer way of contolling the migration down time based on progress. Reducing thread numbers per se is not the main motivation.
pep8 has changed. Again. Version 1.5.1 has even more draconic requirements. We can comply with them, or, include our version of pep8/pyflakes/pylint in our git repo. danken shudders at the thought of copying the code, but having a git sub-module is a reasonable compromise.
- Infra team to take this task (unless someone else is faster)
- Until that happens, one need to use pep8-1.4.6 or --ignore offending errors.
We've been asked to kill the separate vdsm-devel mailing list, and join it into devel@ovirt.org. There's some resistence in the audience, so we'd do it softly.
Next week, I'll have vdsm-devel members mass-subscribed to devel@ovirt. If you do NOT want to be subscribed, please send me a private request.
Then, for several months, we'd see how it goes, and whether we drown in unrelated Engine chatter.
We had a very (too) heated debate about ignoring failures of setDomainRegularRole() in http://gerrit.ovirt.org/24495/ and http://gerrit.ovirt.org/25424.
The pain point is that relying on domain role (master/regular) is faulty by design. We cannot avoid the cases where a pool has more than one domain with a master role written in its metadata.
One side argued that oVirt should be fixed to handle this unescapable truth, or at least enumerate the places where Vdsm and Engine, both current and old, depend on master role uniqueness.
The other side argued that this is not a priority task, and that we should try to "garbage-collect" known-bad master roles as a courtesy to people digging into domain metadata, and as a means to lessen the problem until we kill the pool concept in the upcoming version.
I hope that I present the debate fairly enough.
Dan. _______________________________________________ vdsm-devel mailing list vdsm-devel@lists.fedorahosted.org https://lists.fedorahosted.org/mailman/listinfo/vdsm-devel
Hey all, back from France
On 04/02/2014 11:29 AM, danken@redhat.com wrote:
Vdsm sync call April 1st 2014
- cpopen:
- Toni: there's a versioning mismatch: the version in pypi is higher than the one on https://github.com/ficoos/cpopen
- Saggi: there shouldn't be any code difference
- Yaniv should sync the two.
https://github.com/ficoos/cpopen/pull/23 <- committed 1.3 tag https://pypi.python.org/pypi?:action=display&name=cpopen&version=1.3 <- uploaded latest code
- We use two options that are missing from Python3's Popen: umask and deathSignal.
- Toni to re-send his Python3 port for cpopen, with tests running on Python3, too.
- Saggi to accept it.
- Infra team to suggest missing features to Python.
fromani described his attempts of consolidating the two migration-monitoring threads into one. The motivation is a finer way of contolling the migration down time based on progress. Reducing thread numbers per se is not the main motivation.
pep8 has changed. Again. Version 1.5.1 has even more draconic requirements. We can comply with them, or, include our version of pep8/pyflakes/pylint in our git repo. danken shudders at the thought of copying the code, but having a git sub-module is a reasonable compromise.
- Infra team to take this task (unless someone else is faster)
- Until that happens, one need to use pep8-1.4.6 or --ignore offending errors.
We've been asked to kill the separate vdsm-devel mailing list, and join it into devel@ovirt.org. There's some resistence in the audience, so we'd do it softly.
Next week, I'll have vdsm-devel members mass-subscribed to devel@ovirt. If you do NOT want to be subscribed, please send me a private request.
Then, for several months, we'd see how it goes, and whether we drown in unrelated Engine chatter.
We had a very (too) heated debate about ignoring failures of setDomainRegularRole() in http://gerrit.ovirt.org/24495/ and http://gerrit.ovirt.org/25424.
The pain point is that relying on domain role (master/regular) is faulty by design. We cannot avoid the cases where a pool has more than one domain with a master role written in its metadata.
One side argued that oVirt should be fixed to handle this unescapable truth, or at least enumerate the places where Vdsm and Engine, both current and old, depend on master role uniqueness.
The other side argued that this is not a priority task, and that we should try to "garbage-collect" known-bad master roles as a courtesy to people digging into domain metadata, and as a means to lessen the problem until we kill the pool concept in the upcoming version.
I hope that I present the debate fairly enough.
Dan.
On Wed, Apr 02, 2014 at 09:29:09AM +0100, danken@redhat.com wrote:
We had a very (too) heated debate about ignoring failures of setDomainRegularRole() in http://gerrit.ovirt.org/24495/ and http://gerrit.ovirt.org/25424.
The pain point is that relying on domain role (master/regular) is faulty by design. We cannot avoid the cases where a pool has more than one domain with a master role written in its metadata.
One side argued that oVirt should be fixed to handle this unescapable truth, or at least enumerate the places where Vdsm and Engine, both current and old, depend on master role uniqueness.
The other side argued that this is not a priority task, and that we should try to "garbage-collect" known-bad master roles as a courtesy to people digging into domain metadata, and as a means to lessen the problem until we kill the pool concept in the upcoming version.
I hope that I present the debate fairly enough.
In order to move these two patches forward, how about:
- Limit the usage of the catching-all "except Exception" and replace it with swalling only the expected IO error
- Add a comment about setDomainRegularRole() being called only as a courtesy garbage-collection attempt.
- Conduct a survey on whether migrateMaster is used by anyone. No supported Engine has it, but I there was a suggestion that it was still expected via the command line.
What do you think?
----- Original Message -----
From: "Dan Kenigsberg" danken@redhat.com To: "VDSM Project Development" vdsm-devel@lists.fedorahosted.org, devel@ovirt.org, laravot@redhat.com, nsoffer@redhat.com, fsimonce@redhat.com Cc: lyarwood@redhat.com Sent: Wednesday, April 9, 2014 3:01:31 PM Subject: Re: [vdsm] vdsm sync meeting minutes April 1st, 2014
On Wed, Apr 02, 2014 at 09:29:09AM +0100, danken@redhat.com wrote:
We had a very (too) heated debate about ignoring failures of setDomainRegularRole() in http://gerrit.ovirt.org/24495/ and http://gerrit.ovirt.org/25424.
The pain point is that relying on domain role (master/regular) is faulty by design. We cannot avoid the cases where a pool has more than one domain with a master role written in its metadata.
One side argued that oVirt should be fixed to handle this unescapable truth, or at least enumerate the places where Vdsm and Engine, both current and old, depend on master role uniqueness.
The other side argued that this is not a priority task, and that we should try to "garbage-collect" known-bad master roles as a courtesy to people digging into domain metadata, and as a means to lessen the problem until we kill the pool concept in the upcoming version.
I hope that I present the debate fairly enough.
In order to move these two patches forward, how about:
Limit the usage of the catching-all "except Exception" and replace it with swalling only the expected IO error
Add a comment about setDomainRegularRole() being called only as a courtesy garbage-collection attempt.
Conduct a survey on whether migrateMaster is used by anyone. No supported Engine has it, but I there was a suggestion that it was still expected via the command line.
You don't call masterMigrate directly. It's triggered when you deactivate the master domain.
vdsm-devel@lists.fedorahosted.org