Meeting 11 April 2016, 2pm
Attended: drallen a2brenna ldpaniak
Agenda:
- Timeline
- Remaining work
- Issues to discuss
Timeline
- Looking good for wrapping up by the 13th! (or earlier!)
Remaining work
- Everyone:
- 103620 - assist with HA mysql.cs maintenance documentation : MySQLHAOperationsManual
- Daniel's revamped based on the HA Testing contents; needs additional data (eg, decision tree detail; slave recovery detail)
- from last week: twiki pages start with 'Mysql' - and we should separate pages for ease of finding in emergency (mysql failover) - via twiki 'include' in master document
- 103621 - wrapup and HA mysql.student.cs (phase two) documented plans
- Anthony:
- Questions for Fraser about locations for tools - will resolve by using /usr/local/lib/mysql/bin and close the ticket.
- crontab - will move to /etc/cron.d/mysql rather than master /etc/crontab.
- 103619 - set up monitoring of HA mysql.cs - and sub items -
- status: two checks are written, require Devon's assistance to put into nagios.
- might do an automated backup checking- if time.
- also: could run a checker to make sure pt-table-checksum isn't running more than two copies.
- daniel hasn't followed up on inventory/nagios monitoring being busted- we've never gotten alerts
- needs testing; needs knowing who receives the alerts- set them to alert-list which includes drallen,fhgunn,ldpaniak,helpdesk,lfolland,dlgawley.
- perhaps we need a mysql task-group (specifically not all databases).
- Portscanning currently has port 199 open to the world - SNMP - Anthony wil restrict to nagios.cscf.
- Daniel: as above.
- Lori:
- 103822 - test failover of master/slave/slave asynch HA mysql.cs
- status: we will leave as-is for now, not failing to 106 - bad timing- exams- perhaps in a week and a half, before new term. Perhaps even once a term? on kernel update? no; Dell 730s reboot after kernel upgrade is faster than 5 minutes.
- 103617 - benchmark HA mysql.cs
- can't reproduce Fraser's tests- there are lua scripts missing? See ST 103617... He'll keep working on it; this would be nice to figure out.
- Lori investigated mysql-standard
performance_schema
database - believes it probably doesn't have interesting data to look at.
--
DanielAllen - 2016-04-08