I don't think anyone thumped anything. Scheduler is working on/off (mostly off).
Right now, I'm getting nothing but HTTP errors/timeouts.

---

I'm getting the occasional "can't connect to server" message every now and again but usually most requests get through.

I havn't had any problems uploading and all 3 here are still getting work to download plus for a change I havn't had to resort to a proxy today to get those downloads done.

Cheers.

---

The kitties are bettin', given Matt's recent news and the way things are hiccupping right now, 'She's gonna crash, Jim.'

My God, Bones, what have I done?

:-)

---

David
Sitting on my butt while others boldly go,
Waiting for a message from a small furry creature from Alpha Centauri.

Midsummer on which planet, Neptune? ;).

Well what a wonderful morning here in my part of Aussieland.

I woke up this morning for the first time in about a week where none of my 3 machines had backlogs of uploads/downloads waiting for me to clear.

Cheers.

---

same here, but uploads will not clear. No problem on cricket graph (other than maxed out) but uploading takes longer than running the WUs so upload queue increasing and no new downloads.

---

Scarecrow's graphs show 100,000s of results being returned per hour, i just wish some of them were mine.
The upload backlog continues to grow.

---

Grant
Darwin NT

Yep, the upload server has definitely gone flakey again.

Cheers.

---

Yeh, the Upload server is most definitely off its food, poor thing.
Never mind a few hours that nice man Matt will be back in and he will burp you...

---

Bob Smith
Member of Seti PIPPS (Pluto is a Planet Protest Society)
Somewhere in the (un)known Universe?

what th f?????????more crap

Well, my6 uploads went through okay but the scheduling server shows as down so I can't report. :( Oh well, it's early, they should get things going again soon.

---

PROUD MEMBER OF Team Starfire World BOINC

Not that I have a big check, but could the maintenance they do during the outage be done "live" if they had the power? If yes, what would it take?

Not that I have a big check, but could the maintenance they do during the outage be done "live" if they had the power? If yes, what would it take?

Not really. Part of it is defragging the database and adding new change sets. Doing that while undergoing thousands of hits would mean certain failure.

Steve

---

Warning, addicted to SETI crunching!
Crunching as a member of GPU Users Group.
GPUUG Website

I suspect that Higgs boson is responseble for these problems. :)

---

Goodbye Seti Classic

I suspect that Higgs bozon is responseble for these problems. :)

It is a quarky little buzzard.....

Steve

---

Warning, addicted to SETI crunching!
Crunching as a member of GPU Users Group.
GPUUG Website

Am I wrong, or did you guys not have new servers installed a while ago?
If that is so, i fail to see any improvement.

I suspect that Higgs boson is responseble for these problems. :)

I won't say a word against it -- it pays my whisky bill! Pleased to see that the ATLAS curves today roughly coincide with the CMS ones I had already seen (as a CMS member). Gives some confidence that we are starting to see results that some people have spent two or even three decades working towards (just one decade for me...).

---

Am I wrong, or did you guys not have new servers installed a while ago?
If that is so, i fail to see any improvement.

Just about this time last year. Spent basically this whole time trying to figure out how to fine-tune and optimize the values and settings to get maximum throughput out of these servers. Disk I/O is the major bottleneck.

---

Linux laptop:
record uptime: 1511d 20h 19m (ended due to the power brick giving-up)

If the disk I/O is the bottleneck, maybe it's time to get some SSDs.
They are not that expensive any more.

Well I think there are some SSDs in use already for the cache or log drive. SSDs are too unreliable for sensitive information (the whole limited write cycle issue). Yeah, it's somewhere around 300,000 writes, but for a database such as this.. 300,000 writes on any given sector won't take very long.

The other bottleneck is the database itself. It needs to have certain settings tweaked to be optimized for the hardware, and it's been a long-going process to tweak one value at a time and see what happens.

If you want more information on this ongoing battle, read through the tech news posts since January 2011. There's a lot of mention of what the issues are and what has been done to try to resolve said issues.

---

Linux laptop:
record uptime: 1511d 20h 19m (ended due to the power brick giving-up)