Hi,
You may have noticed that the OpenEHR web site was unavailable for parts
of yesterday and this morning. The full service is up and running now.
There seems to have been a power failure yesterday between about 10:30
and 15:10 yesterday. This was long enough that the UPS initated
automatic shutdowns of the servers in our server room. When the power
returned the servers automatically restarted.
The OpenEHR web site appeared to failed to start up. On closer
investigation many parts had sucessfully restarted (Jira, Confluence,
SVN, etc), however, the Daisy service had not restarted. As the Daisy
provides much of the public facing pages this gave the impression that
the entire server was down.
After investigating the cause of this problem I discovered that daisy
was failing to restart due to an old lock file that hadn't been cleaned
up properly on exit. After manually removing this lock file Daisy was
able to restart without any problems.
There is some evidence that this is a known problem (at least on
Windows), I will look into this further to see if this has been known to
happen on Linux and whether there is a known cure.
Thanks Anthony. It is certainly annoying when it happens, but the power
outages are worse. Any info on whether this situation will improve at
UCL? I am staring to wonder about a second site somewhere else in the UK
or Europe.
- thomas
Anthony Peacock wrote:
Hi Tom,
Thomas Beale wrote:
Thanks Anthony. It is certainly annoying when it happens, but the power
outages are worse. Any info on whether this situation will improve at
UCL? I am staring to wonder about a second site somewhere else in the UK
or Europe.
The power outage this weekend is completely unrelated to the problems
last weekend.
Last weekend the main 11kV supply to the central campus failed, this was
due to a problem with UCL's supplier (I suspect a underground cable was
severed by contractors) and affected the whole area. Unfortunately this
took out some of the network connections between Archway and the internet.
This weekend was a planned outage (although I wasn't given any advanced
warning of it), to carry out work to make the power supply to this
building much more reliable. The main electrical supply into this
building is so old you can no longer get parts for it. Estates has been
keeping it going for about 18 months by robbing parts from areas that
are not in use. Work is currently underway to bring a completly new
feed into the building from our external supplier and connect it to
newly installed modern distribution equipment. The outage last weekend
was part of the first phase of work. I have been informed today that
there will be another outage this coming weekend, when I know the
details I will let you know. Although these current outages are annoying
the outcome will be a much more reliable electrical supply to this building.
Ok, well it is good to know that the current annoyances will serve a purpose...
Anthony Peacock wrote: