[13:26:21] <framawiki>	 !log tools.totoazero moved jobs to stretch, recreated the venv https://wikitech.wikimedia.org/wiki/News/Toolforge_Trusty_deprecation
[13:26:22] <stashbot>	 Logged the message at https://wikitech.wikimedia.org/wiki/Nova_Resource:Tools.totoazero/SAL
[13:35:16] <framawiki>	 !log tools.totoazero moved webservice to stretch, now using kubernetes backend https://wikitech.wikimedia.org/wiki/News/Toolforge_Trusty_deprecation
[13:35:18] <stashbot>	 Logged the message at https://wikitech.wikimedia.org/wiki/Nova_Resource:Tools.totoazero/SAL
[15:51:58] <Steinsplitter>	 can i run a scipt for a few days in a screen? Because on grid the query does not execute (too long) :(
[15:59:22] <valhallasw`cloud>	 Steinsplitter: why would running it in a screen solve that?
[16:00:14] <Steinsplitter>	 valhallasw`cloud: " MySQL server has gone away" i don't have this when running it in as user.
[16:00:36] <Steinsplitter>	 it is a very big query, it takes 3 hours... :( (i optimized it yet!)
[16:02:57] <valhallasw`cloud>	 Steinsplitter: that's... odd. There really shouldn't be any difference between running on a bastion or running on the grid. 
[16:03:20] <valhallasw`cloud>	 are you running the query in a different way (e.g. from pywikibot vs using the mysql command line)?
[16:03:53] <Steinsplitter>	 nope, it is php mysqli 
[16:03:54] <Steinsplitter>	  329831 0.25020 uncat.bot  tools.sbot   r     02/24/2019 15:38:23 task@tools-sgeexec-0915.tools.     1
[16:04:29] <Steinsplitter>	 maybe it is because in the night there is less load on the mysql server. dunno :(
[16:06:32] <valhallasw`cloud>	 Could be, but if I remember correctly the analytics sql servers do not kill any queries. So that would suggest that somehow retrieving the data is slower from an exec host
[16:07:59] <valhallasw`cloud>	 what I would try to do -- if possible -- is to split running the query from any further processing. So have a script run the query and store the results in a file, and a separate job take the file and process it.
[16:08:44] <Steinsplitter>	 i run it on wikireplica-analytics.eqiad.wmnet :)
[16:09:05] <Steinsplitter>	 will try that!
[16:09:11] <Steinsplitter>	 :)
[16:09:46] <valhallasw`cloud>	 if that still doesn't work, and if the result set isn't too large, the query can then also be run on a bastion, as the only resources used will be NFS
[16:10:33] <Steinsplitter>	 okay, thanks. If it dos not work this time i will run it on the devhost :) thanks.
[21:56:08] <Lokal_Profil>	 !log tools.mapillary-commons migrated to Stretch
[21:56:10] <stashbot>	 Logged the message at https://wikitech.wikimedia.org/wiki/Nova_Resource:Tools.mapillary-commons/SAL
[22:13:17] <Lokal_Profil>	 !log tools.wakt migrated to Stretch
[22:13:20] <stashbot>	 Logged the message at https://wikitech.wikimedia.org/wiki/Nova_Resource:Tools.wakt/SAL
[22:59:58] <Lokal_Profil>	 !log tools.slumpartikel migrated to Stretch
[22:59:59] <stashbot>	 Logged the message at https://wikitech.wikimedia.org/wiki/Nova_Resource:Tools.slumpartikel/SAL