[08:13:09] <jynus>	 depooling db1022
[12:06:44] <jynus>	 pv'ing --rate-limit 20m db1022@sqldata.tar.gz --> dbstore1001.eqiad.wmnet
[17:00:01] <jynus>	 recovering db1022 backup from dbstore1001
[17:15:25] <jynus>	 investigating s5 lag on dbstore1002
[17:24:50] <jynus>	 it is replication being blocked at DELETE /* Wikibase\Repo\ChangePruner::pruneChanges  */
[17:25:24] <jynus>	 now getting better
[17:26:39] <jynus>	 not sure why it hit dbstore1002 worse, but probably either reasearch queries or tokudb temporal issue
[17:31:40] <jynus>	 actually, the issue probably came from the master, there was a peak in connections in de: some minutes ago
[20:52:42] <jynus>	 db1022 now in production configuration- but still to recover a lot of lag
[20:53:08] <jynus>	 will repool it tomorrow with a very low load
[20:54:34] <jynus>	 jessie todo: add mysql to it
[20:55:02] <jynus>	 include a mysql(d)?.service config on the package
[20:55:39] <jynus>	 require /etc/mysql directory as it may not be created before it is used
[21:06:59] <jynus>	 there are backups of db1022 on /srv/tmp on that host and dbstore1001, will delete them tomorrow