-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
Hi,
Later tonight we will upgrade the OS on hyacinth (s3/s4/s6 server). This will involve ~30 minutes downtime, but should have no impact on users.
We will also upgrade the OS on willow at the same time, which should require <10 minutes downtime. If there are no problems with that upgrade, we will also upgrade the web servers (wolfsbane & ortelius); but the cluster software makes this upgrade more intricate, so it may not be completed tonight. That will have no impact on users, since web requests will be served by the other server while one is being upgraded.
This maintenance will take place around 12AM UTC on Saturday (11th).
- river.
On 9/10/2010 11:35 AM, River Tarnell wrote:
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
Hi,
Later tonight we will upgrade the OS on hyacinth (s3/s4/s6 server). This will involve ~30 minutes downtime, but should have no impact on users.
We will also upgrade the OS on willow at the same time, which should require <10 minutes downtime. If there are no problems with that upgrade, we will also upgrade the web servers (wolfsbane& ortelius); but the cluster software makes this upgrade more intricate, so it may not be completed tonight. That will have no impact on users, since web requests will be served by the other server while one is being upgraded.
This maintenance will take place around 12AM UTC on Saturday (11th).
- river.
-----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.16 (FreeBSD)
iEYEARECAAYFAkyKUDkACgkQIXd7fCuc5vIJ3QCfWbdwQ+I/pzof66ffUZwklkwa SUYAoLyp7quTM0LCLlTw9vqIQLyhwTnI =/o9r -----END PGP SIGNATURE-----
Toolserver-l mailing list (Toolserver-l@lists.wikimedia.org) https://lists.wikimedia.org/mailman/listinfo/toolserver-l Posting guidelines for this list: https://wiki.toolserver.org/view/Mailing_list_etiquette
Nice. Downtime for my bots.
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
Andrew Steinborn:
Nice. Downtime for my bots.
Yes, there will be under 10 minutes' downtime while the system reboots after the upgrade. This is unavoidable; the only other option is that we would never, ever upgrade the OS or even reboot the server, which seems like a clearly inferior option.
By the way:
willow% uptime 3:42pm up 249 day(s), 9:06, 6 users, load average: 4.84, 4.71, 4.63
One reboot every 250 days doesn't seem like a imposition worthy of sarcastic comments...
- river.
No you don't. Why so sarcastic?
"Andrew Steinborn" megeekz@arandomblog.co.cc wrote:
On 9/10/2010 11:43 AM, River Tarnell wrote:
Andrew Steinborn:
Yeah, I'm getting it.
Toolserver-l mailing list (Toolserver-l@lists.wikimedia.org) https://lists.wikimedia.org/mailman/listinfo/toolserver-l Posting guidelines for this list: https://wiki.toolserver.org/view/Mailing_list_etiquette
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
Hi,
Slight change to this maintenance: we will first test the Solaris upgrade on rosemary, one of the s1 database servers, before upgrading willow. This will push back the upgrade by about an hour, but will not impact users.
- river.
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
This maintenance was completely successfully. There was about 5 minutes downtime of the web server due to a software issue (which is now fixed).
- river.
toolserver-l@lists.wikimedia.org