Riesel Sieve Down Now

[gatekeeper53]

This message started popping up a little while ago




9/13/2006 4:24:06 PM|Riesel Sieve Project|Error on file upload: Project uploads temporarily disabled for load balancing. Uploads will resume around 22:00 UTC.

[Lagu]

2006-09-14 00:46:43||Using earliest-deadline-first scheduling because computer is overcommitted.
2006-09-14 00:47:25||Rescheduling CPU: project resumed by user
2006-09-14 00:47:39||Resuming network activity
2006-09-14 00:47:39|Riesel Sieve Project|Started upload of file riesel-sieve_206464_1_fact.out
2006-09-14 00:47:39|Riesel Sieve Project|Started upload of file riesel-sieve_206464_1_excl.out
2006-09-14 00:47:41|Riesel Sieve Project|Sending scheduler request to http://boinc.rieselsieve.com/cgi/index.cgi
2006-09-14 00:47:41|Riesel Sieve Project|Reason: Requested by user
2006-09-14 00:47:41|Riesel Sieve Project|(not requesting new work or reporting completed tasks)
2006-09-14 00:47:45|Riesel Sieve Project|Scheduler request succeeded
2006-09-14 00:47:45|Riesel Sieve Project|No start tag in scheduler reply
2006-09-14 00:47:45|Riesel Sieve Project|Can't parse scheduler reply
2006-09-14 00:47:45|Riesel Sieve Project|Deferring scheduler requests for 30 minutes and 48 seconds

It seems to be hopeless. I´ve run out of job yesterday and on my Athlon 1001 is 1 task to crunch then is it stop. I have totally 250 wu:s to upload and do to download. I run Einstein on my AMD 64 but every work takes 9 hours.

Lagu

[Nflight]

The 22:00 UTC time has passed one hour and 10 minutes ago, has anyone seen movement?

[Lagu]

My time is GMT and it is difficult to know when 22.00 UTC is.
A question I have:

Say a wu will be done at least kl 10.30.25 sept 14 and I have not upload this WU because of the issue first sept 15. Will this WU be too late and I lose points?

Lagu :?

[gatekeeper53]

Been almost 2 hours and not even a twich. These things hitting the server so often is making my connection seem like the old dial-up days. I wonder if he could up the multiplier on how long the wu's back off after a failure to send. so they only try about every 10 hours or so? Is there any way to ask? I can't even get to the site anymore.

[dAVE]

Lagu, Bryan has turned of the regenerator so your WUs won’t be sent out again when they pass the deadline. This is so that everyone gets their credits no matter how long it takes to send them in. (Edit; Bryan has now said that he intends to keep the regenerator turned off for only "2 or 3 days" until things smooth out)
I see that the project is not sending out anymore work until they sort out the backlog and he asks us not to stay connected and hammer his servers if possible. I’ve gone over to Einstein and just allow BOINC to connect twice a day to report results to Einstein and get new work from them. I wish BOINC had a feature to allow me to suspend all the RS uploads. dAVE

[bryanRS]

dAVE - don't worry, it will be 2 or 3 days after the backlog is cleared, so I figure not until Monday or Tuesday of next week.

Some general (related) info:

OK everyone, here's your 6:45EDT / 10:45 UTC update:

WU file uploads trickled in overnight, and I'm pleased with how things went. The server seemed to stabilize due to the apache controls I put into place. I've made some more changes, and basically, things are starting to catch up much faster. We had about 4,000 files sent in from 04:00-08:00 UTC, and about 5,000 files from 08:00-08:45. Since I'm heading off to work, I'm going to leave the scheduler turned off still, as I won't be here to watch it. I have upped the apache MaxClients to 35, from 20, so that should also help the catchup pace.

I haven't turned on the scheduler yet, although I do have a script set to update the configuration around 15:00 UTC to open up the scheduler again - my hope is that this will work. If the scheduler is still disabled, I'll manually turn it on when I get home.

As long as things are still running when I get home (22:00 UTC), I'll turn on the validator, sit back, and watch things again. We're much closer, and things are stable.

[Empty_5oul]

thanks for the update bryan.

i see some WU are finally getting through as the retry status hits 0. :D

[Strongbow]

Thanks Bryan,

The uploads are working fine!

I know it's sad but I'm looking forward to dumping all those WUs to see how much of a points jump I get!

[bryanRS]

blackheath - it's not sad, I'm quite interested myself, as I have no clue how much work is really stuck out there - my stats/monitoring scripts have all been off to save load.

--------Latest Update--------
OK - my script kind of worked - I set it to check the DB server first, and since the DB server was not communicating with the BOINC server properly, it stopped the changeover. I've fixed the link between the two and they are now communicating properly.

As of 22:00 UTC, WU reservation & reporting reenabled, it seems that over the course of the day everything was uploaded as needed.

We'll see how things go from here....