AP being split? No sign of that on the server stats. All my machines are set for AP only. Have not gotten any WU's in the last 6 hours. Time for another big drink of MB's

Server Status page shows 3 splitters working on the one file, but the work received in the last hour & turn around time numbers certainly don't indicate any AP work going out.

---

Grant
Darwin NT

It would appear there are AP's being split now.......

Well that didn't last long.

---

Either we're having a VLAR storm or something else isn't right as it's been over an hour since my GPU's have got any MB's. :-O

Cheers.

Ah, yeah.
Or something.
09:05 CST was the last time I got any work, GPU or CPU.

"Project has no tasks available" is the only response to Scheduler requests at the moment.

About the same time here when I received my last MB. Ahh, the folly of keeping a low MB queue in anticipation of having APs available by the end of the work day at the lab. I am about to pay the price for that folly as I am out of GPU work and will soon be out of CPU tasks as well.

Ahh, the folly of keeping a low MB queue in anticipation of having APs available by the end of the work day at the lab. I am about to pay the price for that folly as I am out of GPU work and will soon be out of CPU tasks as well

yes know it too well.

Still unable to get new work, and it appears In progress, Awaiting validation, Awaiting assimilation & several other stats have flat lined.
And trying to view Tasks takes ages (if it doesn't time out).

---

Grant
Darwin NT

742pm est last download of new tasks.. none received since; but there's over 438,000 waiting to be sent out?

I just received 20 MBs all at once and the results pages seem to be loading quicker now. I also saw briefly one AP tape being split but I was not fortunate enough to receive a single one.

That brings up another point. Lately it seems one AP tape is being loaded at a time with breaks of no splitting between the tapes. Anyone else notice this and know a reason why it is being done this way?

Edit: I lied. I only received 12 MBs. I forgot I still had a few left in the queue to crunch.

Well the AP splitters out perform the MB splitter on a scale of about 10 to 1. The AP splitters run far ahead of the MB splitters.

Honestly I think they should just run 2 AP splitters and trickle out the tasks rather than doing the big surges of tasks.

If they decide to load the month of data to split when database was down, then fire up all 7 splitter and let er' roll :)

---

If they decide to load the month of data to split when database was down, then fire up all 7 splitter and let er' roll :)

That sounds fairer to me. I have a feeling there are some with slow connections etc. that normally do not get any APs until the initial feeding frenzy dies down. With one tape every now and then you have a feeding frenzy every now and then and it never slows down.

... it appears In progress, Awaiting validation, Awaiting assimilation & several other stats have flat lined.

It appears they haven't flat lined, they just seem to be updating infrequently.
And "Project has no tasks available" is still the response to work requests.
Should run out of GPU work sometime tonight.

---

Grant
Darwin NT

Just picked up 10 WUs on one system. All I need is another couple of hundred...

---

Grant
Darwin NT

... it appears In progress, Awaiting validation, Awaiting assimilation & several other stats have flat lined.

It appears they haven't flat lined, they just seem to be updating infrequently.
And "Project has no tasks available" is still the response to work requests.
Should run out of GPU work sometime tonight.

After a few hours down the pub I see that nothing has changed(other than the number of tasks onboard going down) so I expect my main rig to start hitting up its backup work in the next hour or so. :-(

Cheers.

The cricket graphs show 40-60 MB/s going out to somewhere. Not reaching the users, apparently.

---

Member of the 20 Year Club

The cricket graphs show 40-60 MB/s going out to somewhere. Not reaching the users, apparently.

That traffic may very well just be going back up the hill to the lab as I've seen very little of it at this end since that rise started here this morning with the sun coming up (my time and it's almost sunset now). The tasks pages for my rigs are still extremely slow to load also. :-(

Cheers.

The lab (8_34) doesn't show a corresponding huge inflow of data. (unless I have the wrong cricket for it...)

---

Member of the 20 Year Club

I am about to be totally out of CUDA shortly - Kepler was out about 3:30am, and I7-3820 only has a few left at this time. Haven't gotten any new work since 3am (3 tasks for K) and 2:40am (2 tasks for I7).

Something is very wrong...AGAIN.

---

I can rule out a VLAR storm as well now as my main rig's CPU can't get any of them either. :-(

2 backup projects better be ready for my main rig to come a calling in 2hrs.

Cheers.

It's all just twisted up in knots again...
Figures, being Friday night.
Sent a message off to Matt.
Don't know when he might get it or what he can do.

Meowsigh.

---

"Time is simply the mechanism that keeps everything from happening all at once."

Yeah, something is twisted. Most requests comes with 'Project has no tasks available' -message, maybe one out of 50 request results new tasks, and those are new, not resends.

SSP has problems too, it's updated only once in hour.

And Tasks-page is slow.

All this looks like database issue, it maybe struggling under heavy pressure, assimilators etc...

EDIT: Well, there is Evil in my Tasks -page: State: All (6660) ;)

---

"Please keep Your signature under four lines so Internet traffic doesn't go up too much"

- In 1992 when I had my first e-mail address -

Yesterday's stats finally finished http://setiweb.ssl.berkeley.edu/stats/ so database might do some real work.