Seems like something is missing cause despite having changed the config our
wikimetrics1 instance is still throwing a bunch of errors like:
File "/usr/local/lib/python2.7/dist-packages/sqlalchemy/engine/base.py",
line 238, in _revalidate_connection
"Can't reconnect until invalid "
StatementError: Can't reconnect until invalid transaction is rolled back
(original cause: InvalidRequestError: Can't reconnect until invalid
transaction is rolled back) u'SELECT user.user_id AS user_u
Restating the instance fixes things but somehow it gets "blocked" again. Is
there something special about how these connections need to be managed?
Config is set up as follows (user removed)
SQL_ECHO : False
WIKIMETRICS_ENGINE_URL : 'mysql://<******>@localhost/wikimetrics'
WIKIMETRICS_POOL_SIZE : 100
MEDIAWIKI_ENGINE_URL_TEMPLATE :
'mysql://<****>[email protected]/{0}_p'
MEDIAWIKI_POOL_SIZE : 200
CENTRALAUTH_ENGINE_URL :
'mysql://<*****>@labsdb1002.eqiad.wmnet/centralauth_p'
REPLICATION_LAG_MW_PROJECTS : [ 'enwiki', 'eowiki', 'elwiki',
'commonswiki', 'dewiki', 'frwiki', 'eswiki' ]
REPLICATION_LAG_THRESHOLD : 3 # (measured in hours)
On Mon, Jan 26, 2015 at 5:09 PM, Yuvi Panda <[email protected]> wrote:
> Yes they should!
> On Jan 26, 2015 5:04 PM, "Nuria" <[email protected]> wrote:
>
>> Do these hosts also workfor the centralauth
>> database?
>>
>>
>>
>> > On Jan 21, 2015, at 8:20 PM, Yuvi Panda <[email protected]> wrote:
>> >
>> > the DNAT rules are no longer required. Just connect directly to one of
>> > labsdb1001.eqiad.wmnet, labsdb1002.eqiad.wmnet, or
>> > labsdb1003.eqiad.wmnet. That should work.
>> >
>> > I tried this out on wikimetrics1 instance and that seems to work...
>> > --
>> > Yuvi Panda T
>> > http://yuvi.in/blog
>> >
>> > _______________________________________________
>> > Labs-l mailing list
>> > [email protected]
>> > https://lists.wikimedia.org/mailman/listinfo/labs-l
>>
>
_______________________________________________
Labs-l mailing list
[email protected]
https://lists.wikimedia.org/mailman/listinfo/labs-l
