-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
Hi,
At 05:00 UTC 2016-05-16 an automated component update caused the mirrorlist service for CentOS to go down for all IPv4 based services. The IPv6 service for mirrorlist.centos.org was unaffected.
By 08:11 UTC 2016-05-16 I rolled back the impacted components, disabled the update mechanics, and restarted services.
By 08:15 UTC 2016-05-16 Services were returning to normal.
By 08:23 UTC 2016-05-16 we had multiple confirmations from around the world that services were restored.
- --------------- system wide followup:
- - We will work to extend and add where needed, better testing around each component involved in such roles.
- - I will work with Fabian and make sure that all our automated component and system changes that impact a production service are only run during regular working hours for the team.
- --------------- root cause:
all the components involved in the downtime are backed up and we will start looking at the root cause of why services went down, for now the immediate focus was to restore services which was done by a rollback.
- --------------- reporting issues:
Note that for real time, time sensitive issues always drop into #centos-devel on irc.freenode.net and let us know, along with filing a bug report at bugs.centos.org - For any non time critical issues, please report them at bugs.centos.org against the 'Infrastructure' project, and we will aim to address them as soon as possible.
- -- Karanbir Singh, Project Lead, The CentOS Project +44-207-0999389 | http://www.centos.org/ | twitter.com/CentOS GnuPG Key : http://www.karan.org/publickey.asc
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
Hi,
Apologies for this, the correct date should be : 2016-04-16 Sat. The timestamps remain accurate.
On 16/04/16 10:09, Karanbir Singh wrote:
Hi,
At 05:00 UTC 2016-05-16 an automated component update caused the mirrorlist service for CentOS to go down for all IPv4 based services. The IPv6 service for mirrorlist.centos.org was unaffected.
By 08:11 UTC 2016-05-16 I rolled back the impacted components, disabled the update mechanics, and restarted services.
By 08:15 UTC 2016-05-16 Services were returning to normal.
By 08:23 UTC 2016-05-16 we had multiple confirmations from around the world that services were restored.
--------------- system wide followup:
- We will work to extend and add where needed, better testing
around each component involved in such roles.
- I will work with Fabian and make sure that all our automated
component and system changes that impact a production service are only run during regular working hours for the team.
--------------- root cause:
all the components involved in the downtime are backed up and we will start looking at the root cause of why services went down, for now the immediate focus was to restore services which was done by a rollback.
--------------- reporting issues:
Note that for real time, time sensitive issues always drop into #centos-devel on irc.freenode.net and let us know, along with filing a bug report at bugs.centos.org - For any non time critical issues, please report them at bugs.centos.org against the 'Infrastructure' project, and we will aim to address them as soon as possible.
- -- Karanbir Singh, Project Lead, The CentOS Project +44-207-0999389 | http://www.centos.org/ | twitter.com/CentOS GnuPG Key : http://www.karan.org/publickey.asc
-
Karanbir Singh