LHC Computer Grid - Rollout
> [mailto:[log in to unmask]] On Behalf Of Oxana Smirnova said:
> Right, and then I'll miss the announcement of the GOC DB
> being down, or that there is a new GLUE2 draft 8-)
The downtime broadcasts are quite easy to separate from others, the
format is different, so I probably still see the relevant things - and
if it's RAL, which is about the only one I'd care about, I probably know
it anyway :) Anyway it's completely unfeasible to really read them, the
best I can do is glance at the subject lines before deletion and hope I
might notice something interesting.
> > It certainly makes you realise just how unstable the grid is ...
>
> Not really; neither me nor my fellow VO members would even
> noticed that a site in Kharkiv had unscheduled maintenance of
> power supply feeder unless I've gotten 8 messages about it.
True - one conclusion might be to be impressed at how resilient the grid
is and how little you notice the fact that sites are constantly coming
and going. Unfortunately RAL's rather-too-frequent disappearances are a
little more serious - maybe we need a special channel to announce the
latest power cut/network failure at RAL ... at the very least it's quite
hard to mark services as down when the GOCDB is one of them! (I live
near RAL and I get a power cut about every 3 months, so RAL itself
actually does fairly well - rather ironic given the large power station
only a few miles away.)
> It reminds me of 1994, when NCSA kept a Web page where they
> published monthly lists of *all* new Web sites. There were
> like 8-12 new sites every month, was me favorite bookmark.
Not to mention the original webcam, pointing at a coffee machine in
Cambridge :)
Stephen
PS Can we also mention the fact that every GGUS upgrade seems to
generate about 50 mails?
|