#maria Log v0.1

logs Catalogue

This page loads from the db, refresh as and when. It displays 24 hrs
For older data http://marialog.archivist.info/previous.php, logs go back to 2009-01-18
Some user stats
Table logs
date_time
user
message
2018-01-19 02:47:07
allizom
Hi people! I have a test installation of Fedora 26 where I was learning wordpress, but since when I updated the OS, wordpress has shown "error establishing a database connection". I've run journalctl and it seems mariadb fails to start. /var/log/mariadb/mariadb.log seems to indicate something about "GSSAPI plugin". Pastebins: https://paste.fedoraproject.org/paste/RLwifzNdSVayh-NeBRASPg/raw https://paste.fedoraproject.org/paste/xjdPtgMAmHqf5nVjg5fSpw/raw
2018-01-19 04:47:46
gunix
allizom: first please check that mariadb is listening: "ss -tlnp"
2018-01-19 04:48:42
gunix
you should also check the unix sockets: "ss -xlnp"
2018-01-19 04:49:37
gunix
also please be aware that fedora is a testing distribution that is not fit for production environments. for a stable linux distribution, choose debian, ubuntu LTS or centos/redhat.
2018-01-19 04:50:32
gunix
fedora can still be used for desktop / laptop, but use docker or KVM to provide you with a running environment for wordpress and mariadb. this can avoid compatibility issues that can appear after upgrades, espacially on a desktop/testing distro
2018-01-19 04:51:19
allizom
gunix: it doesn't seem it is listening. And I know that, this is just a test as I said
2018-01-19 04:58:46
gunix
allizom: good, i wanted to put this out just to be sure
2018-01-19 04:59:01
gunix
allizom: try rebooting and restarting mariadb, if rebooting is possible
2018-01-19 04:59:20
allizom
definitively possible. just one minute
2018-01-19 04:59:22
gunix
also check if mariadb process is still up, maybe a ghost is holding lock on the socket
2018-01-19 05:04:38
allizom
gunix: rebooting didn't help it. both systemctl and ps show maria as not running
2018-01-19 05:05:13
gunix
maybe uninstall mariadb and reinstall? this seems like a failed upgrade
2018-01-19 05:05:18
allizom
logs show the same eror as before
2018-01-19 05:05:32
allizom
I can try to reinstall
2018-01-19 05:08:04
gunix
i would ...
2018-01-19 05:08:18
gunix
and if it still fails, submit an issue on github
2018-01-19 05:08:29
gunix
complete uninstall is best
2018-01-19 05:10:09
allizom
gunix, that worked
2018-01-19 05:10:39
allizom
in hindsight, I should have tried it on my own
2018-01-19 05:11:03
allizom
thanks!
2018-01-19 05:12:49
gunix
welcome back allizom (you dced)
2018-01-19 05:13:01
gunix
so reinstall worked? it was a packaging issue related to the upgrade?
2018-01-19 05:13:03
allizom
yeah
2018-01-19 05:13:08
gunix
nice!
2018-01-19 05:13:13
allizom
I don't know for sure
2018-01-19 05:13:49
allizom
thanks!
2018-01-19 05:14:15
gunix
allizom: usually trying all the easy solutions before going into deep troubleshooting is better. i work as a sysadmin and every time somebody comes to me with case i am like "well did you test 1 2 3 4 5 6 7 ? no? well, test all that and if still doesn't work come back"
2018-01-19 05:14:27
gunix
usually they come back only with "yea, it was that"
2018-01-19 05:14:56
allizom
yep, I should make myself a checklist like that
2018-01-19 05:15:01
allizom
good idea
2018-01-19 05:32:42
qwebirc66649
Hey! I have a question on persistpoolmax in maxscale. I have set it to 100 on 3 of my servers and this is reflected in maxadmin show server but the number of connections opened to the backend are far higher.
2018-01-19 05:33:39
qwebirc66649
Im using sysbench to generate 1K connections. Im also using the readwrite splitter, could this be the reason the backend connections arent being restricted to the size of the pool?
2018-01-19 06:23:57
markusjm
qwebirc66649: answered the same question on #maxscale :)
2018-01-19 06:27:02
markusjm
The persistent connection feature of MaxScale is not a connection pool in the traditional sense i.e. it does not limit the number of open connections to persistpoolmax. It will keep a maximum of persistpoolmax connections open even if they aren't used but as sysbench opens the connections at startup, there's no chance to reuse any of the connections. I think mysqlslap would create new connections. Then you can monitor the
2018-01-19 06:27:02
markusjm
counters for connection amounts in the backends.
2018-01-19 06:28:35
markusjm
then there's also the fact that client side connectors also implement connection pools
2018-01-19 06:28:50
markusjm
e.g. JDBC is pretty much always used with something like HikariCP