gevent: bump to >=1.0 #104
Conversation
|
Hi @theanalyst here's the PR for issue #78 |
zhouyuan
force-pushed
the
update_gevent_1
branch
from
0841bd8
to
45f5eac
Compare
|
@zhouyuan Thank you very much! I'll give this a try to see if it solves the DNS issues we're seeing in teuthology-openstack |
| @@ -195,7 +195,7 @@ def main(): | |||
| warmup_pool = gevent.pool.Pool(size=100) | |||
| for file_name in file_names: | |||
| fp = next(files) | |||
| warmup_pool.spawn_link_exception( | |||
| warmup_pool.spawn( | |||
There was a problem hiding this comment.
I know spawn_link_exception is deprecated from 0.13, but is it supposed to be replaced by spawn? (iirc there is something like link_exception)
There was a problem hiding this comment.
@theanalyst
Yes there need to have some exception handling code - this patch was just a quick trying
There was a problem hiding this comment.
@zhouyuan ah ok thanks for taking a look
|
@oritwas should this be against ceph-master or master ? I'm not sure what the difference is ... |
|
@zhouyuan it's a great patch as it will unblock running rgw tests on OpenStack :-) Would you mind adding Fixes: #13885 http://tracker.ceph.com/issues/13885 to the commit message for cross references purposes ? |
|
@zhouyuan oritwas confirmed that master is the target, sorry for the noise |
|
@dachary no problem. I'm still trying to figure out how to add the exception handling code. Did you happen to run into any issue on 'greenlet exception' there or the current code is working fine? |
|
@zhouyuan we did not run into problem. Given that it took about six months to figure this problem out, I suspect that if we leave room for uncertainty in this patch, it will be extremely hard to figure out when and if it manifests itself ;-) Are there tests we can run that are guaranteed to cover all lines modified by this patch ? Note that I'm 100% ignorant of s3test and rgw, that does not help. @theanalyst and @oritwas may have a more informed opinion. |
|
+1 on merging from me; still would be better when the exception handling is added. |
There are some DNS resolving issue on 0.13.6. Bumping to >=1.0 fixes the issues for me. Fixes: #13885 Signed-off-by: Yuan Zhou <yuan.zhou@intel.com>
zhouyuan
force-pushed
the
update_gevent_1
branch
from
45f5eac
to
22edd83
Compare
|
@zhouyuan to summarize http://tracker.ceph.com/issues/13885#note-11 : once this is merged into master, it should be merged into ceph-master and it will be used by http://pulpito.ceph.com/?suite=rgw&branch=master Do you think this pull request is ok to be merged or do you need more time to work on it ? |
ghost
merged commit
|
Running http://pulpito.ceph.com/abhi-2016-04-12_06:24:25-rgw-master---basic-multi/ And http://pulpito.ceph.com/abhi-2016-04-12_06:26:07-rgw-master---basic-multi/ |
|
Both the suites were dead, rescheduled them
http://pulpito.ceph.com/abhi-2016-05-02_01:49:58-rgw-master---basic-multi/
http://pulpito.ceph.com/abhi-2016-05-02_01:47:11-rgw-master---basic-multi/ |
|
Both the suites were dead, rescheduled them again
http://pulpito.ceph.com/abhi-2016-05-23_05:50:44-rgw-master---basic-multi/
http://pulpito.ceph.com/abhi-2016-05-23_05:54:08-rgw-master---basic-multi/
|
|
@theanalyst It looks like these tests died due to timeout when scheduling? the logs are all missing(404) there |
|
@zhouyuan yeah so far none of the tests have actually failed (or passed)as they have never been able to run at all, hoping this run might succeed :) |
|
Rescheduled again: with higher priority |
|
rescheduled only on smithi: |
see ceph/swift#7 and ceph/s3-tests#104 Signed-off-by: Kefu Chai <tchaikov@gmail.com>
see ceph/swift#7 and ceph/s3-tests#104 Signed-off-by: Kefu Chai <kchai@redhat.com>
There are some DNS resolving issue on 0.13.6.
Bumping to 1.0 fixes the issues for me.
Signed-off-by: Yuan Zhou yuan.zhou@intel.com