Hello,
for a kernel-update I hereby announce downtime of sql-s1-rr-a, sql-s2 and sql-
s5-user for tomorrow
FRIDAY, 22:30 UTC.
The downtime of each host will be 20-40min. Not affected will be sql-s5-rr.
sql-s1-rr will be switched to rosemary so no downtime there too, but sql-s1
will be slower. See [1] for details.
Sincerely,
DaB.
[1] https://jira.toolserver.org/browse/MNT-1311
--
Userpage: [[:w:de:User:DaB.]] — PGP: 0x2d3ee2d42b255885
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
Hello DaB!
Hello to everybody!
Since about 2 days I got errors because my bot script started running
on mayapple.toolserver.org. As you might remember, back last year (I
think) there were [1], [2] and [3].
[1] https://jira.toolserver.org/browse/TS-1449
[2] https://jira.toolserver.org/browse/TS-1360
[3] https://jira.toolserver.org/browse/TS-1452
In the course of this you installed a backport of opencv2 (the "2" is
important) to nightshade - even this seems to be noted in the links.
This was done to '/usr/local/lib/python2.6/' and it looks like this is
missing on mayapple which causes my bot to stop working. Could please
somebody compile, copy, symlink, ... this backport to mayapple too?
That would be great! Thanks.
Then another issue - I think (again - sorry but yes! ;) we have some
problem with cron job execution. When looking at my bots (even thought
this is not COMPLETELY clear beacuse of other issue mentioned before)
but also at the lower left of [4] to be precise [5] and compare "Week
6" with "Week 7" till now ("Week 9") then we clearly have some issues.
The first part of "Week 7" may be until mid "Week 8" looks resonable,
but the rest looks "free floating" to me which may point to some load
or other issue... what do you think? Thanks again.
[4] http://munin.toolserver.org/Login/hawthorn/cron_jobs_sh.html
[5] http://munin.toolserver.org/Login/hawthorn/cron_jobs_sh-month.png
Thanks a lot for your help and time and greetings!
DrTrigon
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.13 (GNU/Linux)
Comment: Using GnuPG with Thunderbird - http://www.enigmail.net/
iEYEARECAAYFAlE1FlIACgkQAXWvBxzBrDDgPwCg06mWlZ1hzl0O0B//GPsFAee5
CpcAoKecoBwvb7ZOk01JG6DpPoHWsjd9
=YtdD
-----END PGP SIGNATURE-----
Hello,
Toolserver needs a complete change of the officially assigned IP addresses.
Every accessible service will be involved.
This involves several maintenances that will cause interruptions.
The maintenances will also cause interuptions for OSM.
To see if everything comes up correctly reboots will be necessary on all regarding servers.
I will start with parallel IP setup at the load balancers/head nodes damiana and turnera.
I will do this next Thursday, 28th March 20-22 h UTC.
The next maintenances will involve the login servers, web and mail servers.
They will happen at Friday, 29th March 20 - 22h UTC and Sunday, 31th March 20 - 22h UTC.
I hope everything will work after that.
Cheers
nosy/Marlen
I'm trying to install the Exiv2 library (locally, for just me), but run
into the following error:
theo@willow:~/tools/exiv2-0.23$ make
... [some make jibber-jabber]
/bin/sh: test: argument expected
make[1]: *** [XMPMeta-GetSet.o] Error 1
make[1]: Leaving directory `/home/theo/tools/exiv2-0.23/xmpsdk/src'
make: *** [xmpsdk] Error 2
It looks like this issue—'argument expected'—definitely has arisen before
(example: [
https://mail.gnome.org/archives/garnome-list/2006-June/msg00035.html]) but
I couldn't find a solution. Any enlightenment?
--
*Theo (User:Theopolisme on Wikipedia)
http://enwp.org/wiki/User:Theopolisme<http://en.wikipedia.org/wiki/User:Theopolisme>
*
I was just doing some investigating and discovered that we still have
the cur table (it was removed in mediawiki v1.5) and that the jobs
table isnt visible.
I am currently running scripts to gather data in cross languages during
this process while connected some languages such as no, sl, ro my script
fails due to MySQL server has gone away.
OperationalError: (2006, 'MySQL server has gone away')
Sleeping for 19.0 seconds, 2013-03-26 08:25:27
Sleeping for 22.0 seconds, 2013-03-26 08:25:46
Sleeping for 12.0 seconds, 2013-03-26 08:26:09
Sleeping for 5.0 seconds, 2013-03-26 08:26:21
My solution was to put a try except to handle this and reconnect to the
server if the exception error code is 2006 but the problem is I have many
query that I execute in multiple methods. I think an easy way to do is edit
the mysql configure but obviously thats not an option. It doesn't happen
all the time. Any suggestion or solutions? I am using python as my coding
language.
Here is a rough idea of my solution:
except MySQLdb.OperationalError, sqlEx:
if sqlEx[0] == 2006:
logging.info("Caught the MySQL Server Gone away exception
correctly")
logging.exception(sqlEx)
time.sleep(10)
self.connectServer()
self.dbCursor.execute(botUserQuery)
Thank You,
--
Anuradha Uduwage (Anu)
This is slightly outside the scope of this mailing list, But could
someone write a web tool that someone can pass a filename, lang,
project and the script then gets the associated image, lets the user
modify the size (and has a few pre-configured resolutions including
0.1MP ) then download it so it can be re-uploaded (or uploaded via
upload by url if the wiki has it enabled.)
I know this would help a lot of users.
Hello all,
it was discovered that an update at the WMF-side breaks our replication and
makes some manual work necessary. I have run these manual step on the new
setuped s1-server now. The problem is that the other sq-server (rosemary) has
~17k seconds replag and so the problem will occur during my sleep-phase. So
most likely the replication of enwp will stop tomorrow morning and stay so
until a root is awake.
Just as information.
@Nosy: If you are around when the problem occurs: Skip the queries that update
the mysql.XXXX-tables (only these!) and run mysql_upgrade when it complains
about a broken mysql.proc-table (the first should be an update in
recentchanges).
Sincerly,
DaB.
--
Userpage: [[:w:de:User:DaB.]] — PGP: 0x2d3ee2d42b255885
Op 22-3-2013 20:15, Maarten Dammers schreef:
> Op 22-3-2013 10:43, Daniel Kinzler schreef:
>> On 22.03.2013 10:06, Marlen Caemmerer wrote:
>>> Hello,
>>>
>>> Toolserver needs a complete change of the officially assigned IP
>>> addresses.
>>> Every accessible service will be involved.
>> Wow, best of luck :/
> Yeah, good luck. But why are you doing this exactly? I assume ipv6
> won't be affected?
I see that I was on the wrong mailinglist and that Wikimedia got a new
/22 (ip) and /32 (ipv6) from RIPE. That explains it....
Maarten
