Planned Outages

**russellw** · 24-03-2016, 08:04 AM

Good morning

We are in the process of migrating AFF to new server hardware which will be a two step process involving two outage windows.

The first of these will be from 01:30 Saturday March 26th until about 04:30 Saturday March 26th.

During this outage one of the hard disks will be removed from the current server and added to the new one for cloning purposes.

Note: During this period there is some risk to service continuity because removal of the drive necessitates breaking the current Hardware RAID and leaves us reliant on a single disk drive but it's an acceptable level of risk.

The second step will be scheduled at a later date after the cloning and testing of the new server has been completed but we are currently planning the second outage window for the period from 01:00 Monday March 28th until about 05:30 Monday March 28th.

During this outage the database on the new server will be updated to reflect changes made since the clone and the IP addresses swapped between the old and new servers so that DNS continues to work.

We've had more than 5 years use out of this kit so it's more than done what was expected of it and the new server will provide improved performance and greater resilience to failure.

Best regards
Russ

PS: There will now be a 3rd step as we have discovered that one of the hard drives in the current server has failed. There will also be an outage at 01:00 tomorrow (Friday) morning to replace that hard drive. This should take no more than 90 minutes allowing for a backup to be taken before that work commences.

During this outage window we will be replacing a failed hard disk in the server. There will be a degradation in system performance for approximately 12 hours after this is done as the RAID array rebuilds itself.

**supershifty** · 24-03-2016, 11:49 AM

Do you need any "testers" for the migration Russell?

I can break my PC with no ability whatsoever, so the AFF server shouldn't be a problem.

Can make myself available on both mornings if needed.

Cheers!

**russellw** · 24-03-2016, 10:05 PM

Thanks I probably will. At this stage UAT on the new server is scheduled for around 15:00 Saturday afternoon.

Cheers
Russ

Geez Louise · 24-03-2016, 10:41 PM

I will be online Russ. Hope it all goes well 😀

-------------------------------------------------

Just a 70's girl in a confusing world!

**russellw** · 25-03-2016, 03:30 AM

Good morning

The first phase is now completed successfully. The outage was about an hour longer than planned (01:00-03:25) and as previously noted there will be some slowing of server performance for the next 12 hours or so while the RAID array is rebuilt.

Regards
Russ

**russellw** · 26-03-2016, 04:49 AM

The second phase is now completed successfully.

Regards
Russ

**russellw** · 26-03-2016, 06:26 PM

Thanks to a classic piece of stupidity from me, we are probably going to have to delay the 3rd phase as I have rendered the new server unbootable. It is currently being rebuilt but that means all the work done today in terms of installing the 130 odd packages; copying the 50 Gb of data and massaging the configuration files has been lost and will have to be repeated.

I don't have much time available tomorrow so that work will probably have to be completed on Monday and the cutover deferred until later in the week after testing has been completed.

Cheers
Russ

**russellw** · 27-03-2016, 07:21 PM

So, I've managed to get back to almost the state we were in before my mistake yesterday and now the copy of our data will commence. There may be some small performance hit during the copy process.

Cheers
Russ