Difference between revisions of "Admin TODO"
Jump to navigation
Jump to search
| Line 22: | Line 22: | ||
***i think sirius is missing some crontab to get the svn and generate certain files | ***i think sirius is missing some crontab to get the svn and generate certain files | ||
*** need to create dns entries for {www,images,download}{1,2,3,..} and matching vhost | *** need to create dns entries for {www,images,download}{1,2,3,..} and matching vhost | ||
| − | **prepare "maintenace" pages for www, trac, wiki, forum | + | **for when sh**t happens: |
| − | **prepare simplified pages with only a few links | + | ***prepare "maintenace" pages for www, trac, wiki, forum |
| + | ***prepare simplified pages with only a few links | ||
* Mail | * Mail | ||
**create an admin ml | **create an admin ml | ||
** make sure the cron send to admin@ if and only if something bad happened | ** make sure the cron send to admin@ if and only if something bad happened | ||
Revision as of 23:53, 25 March 2009
Various stuff to do on the servers to improve services or relibility
- upgrade Albiero and sirius to Lenny
- Imply a reboot to change kernel => beware of eth0 vanishing
- physical access ?
- Hardware
- Switch gigabit between the eth1 of ganesh, albiero and sirius
- way to get albiero and sirius console from ganesh
- way to have ganesh console on albiero (iLO *does* fail)
- Database
- move them all to albiero
- move them all to pg 8.3 (requires lenny)
- do we need/can have a fall back if albiero is down
- Web
- optimize pg request
- make sure the web work even if the databases are down
- important for fastcgi (maybe tweack various timeout)
- avoid ugly error on apache
- mirroring/reverse-proxying/round-robining
- i think sirius is missing some crontab to get the svn and generate certain files
- need to create dns entries for {www,images,download}{1,2,3,..} and matching vhost
- for when sh**t happens:
- prepare "maintenace" pages for www, trac, wiki, forum
- prepare simplified pages with only a few links
- Mail
- create an admin ml
- make sure the cron send to admin@ if and only if something bad happened
