Subject[dq] List-server / web-server outage
FromMartin Kealey
DateSun, 14 Oct 2007 09:29:45 +1300 (NZDT)
Suad and I offer our belated apologies for the outage last week; our main
server in Auckland failed late on Tuesday during a thunderstorm. This
disrupted your mailing lists and web service.

I became aware of the problem at about 6:30 on Tuesday, and phoned the data
centre (just before they closed) and asked them to reboot it. By the time it
was evident that this hadn't succeeded they were closed for the night.

I made further attempts in person to reboot it on Wednesday (unfortunately
not until after lunch as I had a crisis as my "day job" and Suad was in
Melbourne last week) and then decided to removed it so I could copy its data
onto a replacement server. The new server ready by about 7pm and was
installed about 8am Thursday when the the data centre reopened.

I recommend that if you have mailing list subscriptions with other services
they should be checked in case bounces have triggered automatic
unsubscription (this is unlikely but possible).

The outage was lengthened and exacerbated by several factors:
 - a crisis at my day job meant I wasn't able to take time off to deal
   with it (and I was on the point of calling in sick to my main job if
   there hadn't been a crisis);
 - Suad was in Melbourne for the week, so nor could he;
 - the data centre had recently changed its security policy, and it took
   an hour or so to sort out my access;

On the plus side, all user data was recovered from the active media (so no
"time gap") and as far as I know all software has been reinstated. However
we're only human and might have overlooked something, so please let me know
if anything is missing.

I'm currently negotiating for a "fail over" server at another location and
will let you know which option we choose.

And again, our apologies for the loss of service.

-Martin


-- to unsubscribe notify mailto:dq-request@dq.sf.org.nz --