Difference between revisions of "Sysadmin/Announcing downtime"
| Line 22: | Line 22: | ||
The alias "'''hosted''' @ '''sugarlabs.org'''" forwards outage notifications to several | The alias "'''hosted''' @ '''sugarlabs.org'''" forwards outage notifications to several | ||
individuals and mailing lists who have services hosted at Sugar Labs. The alias is handled | individuals and mailing lists who have services hosted at Sugar Labs. The alias is handled | ||
| − | by our MX, therefore it won't work if [[Machine/ | + | by our MX, therefore it won't work if [[Machine/sunjammer]] itself is down. |
Use this to notify about ahead-of-time about planned maintenance and | Use this to notify about ahead-of-time about planned maintenance and | ||
Revision as of 18:37, 30 May 2011
How to announce service outages
identi.ca
Parties interested in receiving these notifications subscribe to our Infrastructure_Team/Contacts#Services_status feed.
To send up/down notifications for hosts and services, send messages on [1] crafted like these
!sugarlabsstatus {wiki,www,planet,lists}.sugarlabs.org: maintenance downtime for solarsail in 10mins
!sugarlabsstatus {wiki,www,planet,lists}.sugarlabs.org: solarsail going down now
!sugarlabsstatus {wiki,www,planet,lists}.sugarlabs.org: back up
!sugarlabsstatus beamrider: machine is unreachable, pls contact co-lo admins
Please do not abuse this service, as some users configure their account to forward notifications as text messages to their mobile phones and may have to pay for each SMS received.
hosted@sl.org
The alias "hosted @ sugarlabs.org" forwards outage notifications to several individuals and mailing lists who have services hosted at Sugar Labs. The alias is handled by our MX, therefore it won't work if Machine/sunjammer itself is down.
Use this to notify about ahead-of-time about planned maintenance and possible service outages.
Here's an example of what a notification may look like:
Subject: develer.com: switch to new uplink, Aug 13 23:00 UTC The FSF sysadmins are going to relocate our virtual machine Sunjammer today in the evening. The new host has faster processors, twice as much memory, a faster RAID1 array and generally better hardware. The downtime is projected to last 15-30 minutes, but it may last longer in case of unexpected issues. Apologies for any inconvenience, -- Your dedicated BOFH
Please format the subject line like in the example, provide detailed explanation of what services are affected, along with an estimate of time required to complete the work in average and worst-case scenarios.