04/08/2003 - 04/09/2003 Power Maintenance/Modifications

[porcupine]

Hello Everybody, Here's the scoop:

The Power plant that the APC Masterswitch units APC1 Through APC6 Operate off are soon to be decommissioned. Since our arrival in 151 Front in November, we've pushed this power plant to the limits of it's current capacity, and as a result, a new, signifigantly larger power plant has been rolled out, and the old one will be decommissioned.

As a result, APC1 Through APC6 (and all attached clients) will be relocated to the new power feeds between the hours of 11:00pm on 04-08-2003, and 07:00am on 04-09-2003. Priority Colo Staff will fully manage this, and will do soft shutdowns on every server before shutting the APC's down to relocate the power to the new feeds. The APC's will be shutdown and brought up one at a time, we will ensure VIA our port monitoring that every server, and every port thats monitored is back up before we move on to the next.

Priority Colo will *not* be held liable for any servers that cannot be properly shutdown due to lack of proper root passwords. If we cannot login to a server, we will make best efforts to properly shutdown VIA console.

Expected downtime is approximately 5 minutes per MasterSwitch. If we have your current root password on file, you have little/nothing to worry about and this will be completely transparent.

Shared hosting customers on www3 through www6 will also be affected by this. Dedicated and Colocation customers will be notified VIA email also, this thread merely serves as a reference.

Regards,

[porcupine]

Additionally I'd forgotten to note.

This is being done tuesday as the yearly maintenance for the building is being done on the 12th (Saturday). During this time all building power is shut off for 12 hours, and customers rely on their UPS and Generators to stay up. Obviously we'd prefer to be on the new UPS, Generator systems, etc. with higher capacity, "just in case".

Incidentally, if the email that was sent out hit anyones pagers at this Un-Godly hour, please forgive me, we just realised it might have strayed a few.

[porcupine]

Ok guys,

Well for the most part it was fairly smooth... There was a dozen or so boxes which had problems (services not in the startup, one or two with unbootable kernels selected, and a few with mysteriously non-working services [eg cpanel apache, which we merely recompiled to repair]). All of the APC's were migrated, except APC's 5 and 6, which blew the new circuit we put them on in error (they were put on the wrong one accidentally. As a result we'll have to re-schedule for the final transfer of apc's 5 and 6. We may move them this saturday night, we may decide to move them in a month (depending on how long we have until the current plant is turned down).

We will be re-contacting everyone on those two apc's at the first opportune time once i've re-scheduled their moves.

Regards,