| jamespage | icey: I've MIR reviewed all of the octavia packages (and associated depends). Majority need security review as well (assigned to security team) and there are a couple of non-blocking recommendations as well | 06:10 |
|---|---|---|
| icey | cool - how would you feel reviewing the sqlalchemy-utils one? | 06:11 |
| icey | :-D | 06:11 |
| icey | oh jamespage - looks like you also did that already, thanks :) | 06:16 |
| jamespage | I did | 06:16 |
| skylite | I just checked my mysql status with systemctl and I see: Status: "InnoDB: Error: Database page corruption on disk or a failed file read of tablespace..." I wonder if anyone has a tool that sends email alerts when something like this occurs? | 09:56 |
| skylite | are there any good tools for that? | 09:57 |
| RoyK | skylite: I use zabbix for that sort of things, but I would guess systemd has some setting for it somewhere? | 10:43 |
| skylite | RoyK: Im looking into it in #systemd it's not that easy by default | 10:44 |
| skylite | but thanks | 10:44 |
| RoyK | skylite: OnFailure=notify-failed@%n | 10:50 |
| RoyK | not that? | 10:50 |
| RoyK | https://serverfault.com/questions/694818/get-notification-when-systemd-monitored-service-enters-failed-state/701100 | 10:51 |
| skylite | no this is only for notifying if the service is down | 10:52 |
| skylite | for example my mysql is not down I just have a corrupted table for instance | 10:53 |
| skylite | I would see that in the StatusText | 10:53 |
| RoyK | I see | 10:55 |
| RoyK | well, I would just write a zabbix check for that if there isn't anything available already, but then, you'll need to setup zabbix first which my be a bit overkill for monitoring a single machine | 10:56 |
| skylite | I'm using datadog I'll check if they have a solution for this one | 10:57 |
| skylite | I already have a command that gives me the last StatusText back for a service | 10:57 |
| skylite | If nothing I'll write a shell script that alerts me on change | 10:57 |
| RoyK | should work | 11:06 |
| RoyK | or just pipe it though sed or awk or perl or python or something to look for bad signs | 11:06 |
| RoyK | skylite: something like this: if $(systemctl status mysql | grep -i error > /dev/null 2>&1 ) ; then echo No error ; else echo FAIL! ; fi | 11:10 |
| skylite | sure but I'm thinking I want to somehow save the last status and alert if it changes | 11:11 |
| skylite | I feel like I need to know everything for now :D | 11:11 |
| RoyK | skylite: you may be able to get some ideas from this https://share.zabbix.com/operating-systems/systemd-service-monitoring | 11:22 |
| RoyK | it's way broader, though, but possibly worth a look | 11:23 |
| RoyK | skylite: erm - sorry - forget about that - it just shows status, not details | 11:26 |
| RoyK | it'd be nice i systemctl could output its status in some parsable format like json or something | 11:29 |
| skylite | I bet it can | 11:29 |
| RoyK | couldn't find anything in the manual | 11:30 |
| skylite | busctl get-property org.freedesktop.systemd1 /org/freedesktop/systemd1/unit/mysqld_2eservice org.freedesktop.systemd1.Service StatusText | 11:30 |
| skylite | this gives me back only the StatusText string | 11:30 |
| skylite | for mysql | 11:30 |
| RoyK | should be ok to script then | 11:41 |
| === klokken is now known as iopl | ||
| === iopl is now known as klokken | ||
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!