Panic Mode On (9) Server problems

From Matt's post on the Technical News section..

"We're also fairly pegged at our network limit again, I think thanks to the workunit turnaround time being pretty low (i.e. fast). Plus I have to send extra raw data to our archive over the same link. Oh well. Expect data transfer headaches for the next qwhile.
"

I think that we are still in the 'Qwhile' period, and catching up after the extended Tuesday outage is bogging down the network link.

No servers are down, there are no faults to fix. Just cache a couple of days work and you wont have any problems. The system is tollerent of temporary faults and network overloads and will sort itself out in due course.

Ian

There was prior warning here and the front page still says,
[quote]Weekly Outage
Every Tuesday morning (Pacific time) we have a 3-4 hour outage for database maintenance. The upload/download servers and some web pages are offline during this time. Due to extra work today's outage will start earlier and end later than usual.

did not know i was in court, hope the verdict is "you are free"

OK, if you're done, let's start a new panic: the Server Status page hasn't updated since "As of 15 Oct 2008 6:00:20 UTC".

And if anyone does get out of bed at 2 a.m. local time, I expect them to concentrate on fixing it, not transcribing what I've just written to the front page.

(And when I say 'expect', I mean that's my prediction of what they will do, not any sort of demand on my part).

OK, if you're done, let's start a new panic: the Server Status page hasn't updated since "As of 15 Oct 2008 6:00:20 UTC".

And if anyone does get out of bed at 2 a.m. local time, I expect them to concentrate on fixing it, not transcribing what I've just written to the front page.

(And when I say 'expect', I mean that's my prediction of what they will do, not any sort of demand on my part).

Sorry that You don't understand whaT I mean. What I mean is that after serveral years of ignoring to write the shortest explanation possible, it is starting to irritate me. That is what I'm talking about. This total lack of care.

Calm down Tonne.

Your point is noted.

The crew is working hard to solve the problem(s). They do what they can. We are all humans and can fail in our work.
____________

Besides you, "Who reads the front page"?

What is it You don't get. 2 lines on the front page. 1 minute work. I'm not talking about writing a bestselling novel or e-mailing every user individually.

X server is down because of X and we expect to be up agian in X days/hours

How long can that take to write on the front page?

Besides you, "Who reads the front page"?

I asked the question, "Who reads the front page"? because a few years ago a notice was put on the front page about two days before a known outage.
Within about 2 hours of the outage occurring there were about 10 threads started here, the cafe and on Q&A wanting to know what was going on.

And as obviously those people didn't and I admit I very rarely do, then "Who does read the front page"?

Also observing family and friends when they go to frequently visited sites the cursor is usually hovering over and clicking next required page before front page has fully loaded.

We've all - myself included - had a bit of fun at Tonne's expense, but I think he deserves a serious response as well (even in what is meant to be a light-hearted thread). Actually, I do understand what he means, and I have felt - and posted - similar opinions myself.

Let's examine the fault-handling process, from the staff's point of view.

1) First possible indication of fault - user input
2) Staff become aware of problem
3) Potential reporting point
4) Staff diagnose fault, form plan of action
5) Another potential reporting point
6) Staff fix fault
7) Final potential reporting point

There is no way that (1) is going to find its way onto the front page - UCB isn't going to allow access for outside volunteers, however well vetted, to modify the home page. And the staff can't do it, because at this stage (early hours of the morning? weekend?) they aren't aware of the problem.

There is the possibility of amending the home page at line (3), but I don't think many of us are asking for that. Tonne, in particular, doesn't want reporting at this point, because in two (edit - now three) separate posts he's indicated that he would like an indication of the expected duration of the outage: and that requires diagnosis first.

So the earliest possible useful front-page notice comes at step (5). But by this stage, they've already got the closet doors open, or the server logs in a terminal window, or whatever: it's actually much more efficient to apply simple fixes at that moment, with the tools to hand, than to move to a completely different operating environment to modify the home page. Which means we're at line (7) already, and personally I'm happy for the post-mortem to be conducted in Technical News, rather than the front page.

Of course, the weasel word in that last paragraph is "simple". The obvious counter-example is the day the motherboard died in Thumper Mark I. Because that was a pre-production, prototype server, it couldn't be fixed with a simple switch-out, and the project was down until a new server could be built and shipped trans-continent. That sort of outage deserves, and I seem to remember received, regular updates on the front page.

I don't know why sometimes some people are angry to the project..

We all are long enough here to know that sometimes the project have problems.. and?

Rise the cache of WUs.. and everything is fine.. where is the problem??


EDIT:
BTW.
(200. post in this thread.. ;-)
____________


>Das Deutsche Cafe. The German Cafe.<