So.... a few users couldn't print on Monday afternoon. 'How odd', I thought and tried to remote in to the print server. RDC doesn't work, so off i go to physically look at the beast, and find it powered down. Hmmm... Switch it back on and it springs in to life and works perfectly as before.
Later investigations show that it shut down with no warning - nothing mentioned in the various event logs until after it was brought back up, then 'previous shutdown was unexpected'. It gets stranger, however...
Print server sits next to, and shares a UPS with, our new Curriculum server. Curriculum server holds all staff and pupil files, and several shared drives. Curriculum server also shut down at the same time as the printer server; luckily it restarted itself as well, so disruption was minimal. Same story with event log; no mention of a shut down until after it has restarted.
Curriculum server holds the UPS software for the pair. When a disruption of mains power is detected, server is configured to stay running until extent of battery runtime, then shut down. Before it shuts down, however, it uses a cmd script to send a shutdown command to the print server. Though this might explain both servers shutting down, there is no record in the UPS software or Windows event log of any power events, etc. Also our Admin server, which is connected to another UPS and which stayed running, recorded no power events.
The REALLY odd thing about this is that this is the third time its happened. Each event has been a monday, between 13:59 and 14:05. Not every monday, mind, alternate mondays. As we run a two week timetable, this corresponds to Week A, and a period where the ICT room that contains the server cupboard is free for staff to book it for their classes. One particular Maths class has been in the room on every occasion.
Not sure what to make of this - although the same pupils are in the ICT room everytime the servers restart, I can't see how they could cause the issue. The server cupboard door is locked at all times and i can't see anything outside that would do it.
if it's affecting both servers connected to a UPS, I would say the first place to look would be the UPS. It may be that one server is better at coping with brownouts / momentary blackouts than the other. Do you have a spare UPS to try? or another one to connect in series?
As for happening at the same time each fortnight... it's not on the same ringmain as something like the D+T department using a welder, or anything else that would suddenly change the mains voltage slightly is it?
(here in the wilds of south Cumbria where the electricity frequently comes in with lumps in it I often have UPSs on my UPSs)
Last edited by BatchFile; 25th November 2009 at 09:33 AM.
No spares, unfortunately... The logging on the UPS connected to the unaffected server shows voltage records - as far as i can see there were no brown-outs or other events that might explain things, unless the cause is in school and effecting mains sockets in the same little room differently
The timing puts it towards the end of a lesson, so its not like it could be a class all switching their PCs on at the same time..