=== Toushirou is moving ===
As you may know, Sivit, one of our sponsor, was acquired by Nerim, so
we have to move this machine in another location (Equinix). Tomorrow
evening, Hivane folks, Yok, and me, will work on this move, and
Toushirou will be down a short time. Nothing critical would lack, but
a few services will be unavailable (mostly web related, except
webmail/webdesk which is elsewhere).
=== More memory ===
Toushirou has now more memory (4GB instead of 2) since 2010-05-29, and
is working fine. As our stability problem seemes to be memory bound,
it is probably a stopgap, but anyway this machine needed more, so this
is a good and not so expensive evolution.
=== Recent problems ===
* 2010-06-04 ~11:10: due to electrical problems in the bay, Orfeo was
down during about half an hour.
* 2010-12-06: the main Elwing's disk died, so it was quite a pain in
the ass, as this machine had no RAID yet, but we managed to repair it
quickly; nevertheless the disks or something in the OS is not working
well, and we have horrible performance. It should not affect DC's
services too much, as most of them moved or are duplicated elsewhere.
* the second repaired disk for Daneel died after less than a minute of
stress when rebuilding the RAID; happily the spare disk was here to
take it over again until a new disk exchange is done...
Hope this disease affecting disks will soon be eradicated...
=== Ruby for All ===
Recently, we removed the badly working mod_ruby apache extension,
stopped using fcgid which was working but not so efficient/responsive,
and switched to mod_passenger, which seems to behave quite well. You
should be able to activate Ruby apps you need quite easily now, in
your vhost, or in your personnal area (not limited to a single app).
You can read more on this subject here:
and do not hesitate asking us if you need to override the ruby-app
detection in your vhost, or need specific settings.
Marc Dequènes (Duck)