At this time yesterday I had 261 waiting to upload. Today I have 267 waiting. During the planned outage I got up to nearly 400, so there was some progress, but that seems to have stopped.

---

I thought so too... I had over 8 with Xeon-32a, but this morning I was down to 4 ... hoping ... well ... so far none of the others ... oops ... I lied ... it looks like 2 more went up ...

---

The cricket graphs are back to what they were before the outage, and in some cases, lower. But am still only getting 3 or 4 WU's uploaded / day, so my queue is growing.

It's been around 3 days since the outage... you would think that it would stabilise by then.

Oh, and I got shot at tonight while having dinner at my local restaurant. Only one .22.

---

My observation on those graphs is it needs to be nearer 60M/bps than 50 to be working normally. So I guess things aint right yet.....but they start their day soon so perhaps they will leap straight to it!

I feel fairly sure this is not just dealing with a back log that is causing the problem...there is something else not right.

---

Very sloooow up/download - I thought it wuz just me.

Is the server status page getting updated?. Looks out of date - or is it me

Not able to upload here either. Lots of units waiting, but at least i am getting new ones downloaded. On the other hand, in 48 hours several will be in default, unless I can upload them.

Nick

---

To day they have try to solve this problem for 4 days an still no success.
What kind of university is this? ;)

---

To day they have try to solve this problem for 4 days an still no success.
What kind of university is this? ;)

What problem? There is no problem to be fixed.
OTH, If you are talking about the lack of capacity to handle the workload, then it is simply a matter of budgets - it's a university like every other in the world, without infinite money available.

---

But Mike we have both said in the past when the classic user workload comes over it will struggle to recover. They have started I noticed to come with an upsurge in users recently. We were told that this will not affect performance.....many gurus on this board have said this albeit I could not readily agree.....but I think we see the consequences now with these huge delays.

Or is it a different problem altogether? We know not but I feel sure recovery times will get longer and longer until some additional processing power is brought to bear.

Close classic now I say and get that hardware re-deployed ASAP. Without doing this new users will walk for certain!

---

But Mike we have both said in the past when the classic user workload comes over it will struggle to recover. They have started I noticed to come with an upsurge in users recently. We were told that this will not affect performance.....many gurus on this board have said this albeit I could not readily agree.....but I think we see the consequences now with these huge delays.

Or is it a different problem altogether? We know not but I feel sure recovery times will get longer and longer until some additional processing power is brought to bear.

Close classic now I say and get that hardware re-deployed ASAP. Without doing this new users will walk for certain!

I think a big part of the problem is they're using obsolete systems. I had a look around at the sun site and after some searching found info on all but the Sun D220R (which they have the most of).

A big problem is that the BOINC UL/DL server seems to favorize client downloads (upload from Berkeley's view), so that client uploads gets a lower priority.

I'd prefer an equal priority, or even a slight favorism on client uploads.

A big problem is that the BOINC UL/DL server seems to favorize client downloads (upload from Berkeley's view), so that client uploads gets a lower priority.

I'd prefer an equal priority, or even a slight favorism on client uploads.

I disagree.

Client downloads correctly get priority to keep hungry hosts fed with new work as soon as possible. Then the results can be returned as a lower priority whilst those hosts are chewing on their new work.

If you're pushing too close to the deadline, then perhaps you've been too greedy with your cache settings... I think the limits are 10 days max which gives 4 days to get the results back worst case.


Also, there's been a tech note saying that their download/upload server is 100% CPU bound...

Happy crunchin',
Martin

---

See new freedom: Mageia Linux
Take a look for yourself: Linux Format
The Future is what We all make IT (GPLv3)

To day they have try to solve this problem for 4 days an still no success.
What kind of university is this? ;)

What problem? There is no problem to be fixed.
OTH, If you are talking about the lack of capacity to handle the workload, then it is simply a matter of budgets - it's a university like every other in the world, without infinite money available.

The problem that you can't down or upload an they have worked on this problem for 100 hour. I certainly expect more form a university. It's not a matter of budget it's a lack of skill.

---

It's all a question of how much patience you have, Tomas. As said, and as far as we know (read: are informed about), there is no problem to be fixed, they aren't trying to fix things. It's just congestion since a lot of people try to upload their stuff.

I can download fine. The couple of problems I may have are probably more due to the development version of Boinc I use.

---

The problem that you can't down or upload an they have worked on this problem for 100 hour. I certainly expect more form a university. It's not a matter of budget it's a lack of skill.

Ever tried to push a watermelon through a keyhole...see how long it takes you...

there is no problem to be fixed, they aren't trying to fix things

LOL :)))) Are you a politician?

---

..... there is no problem to be fixed, they aren't trying to fix things. It's just congestion since a lot of people try to upload their stuff.

I think there is a problem to be fixed. Have said so many times and always get shouted down by the high priests and gurus that have time served. Well let me say it again:

The problem is they will not keep a decent user base unless they demonstrate some level of acceptable capacity and some level of acceptable reliability. Now I have heard the cries from here already.......but I think it needs to be said over and over.....its the people who process who are impportant and if they cannot be convinced the work they do is getting back and credits given then they will clear off. UCB PR is zilch apart from the odd post to keep the natives semi-happy. Even if they can't provide the capacity they should handle folk better than this. Well the PR plain sucks. The whole capacity issue needs to be dealt with, classic closed down to free up hardware and free up bandwidth and a decent PR strategy put in place. As I have said somewhere else during this non productive debacle all the priests and gurus have laughed behind their hands when I say the system is in danger of not recovering from an outage. I suggested that a model should be built to establish what the crucial cut-offs might be so that a decent plan could be made to cope. Of course the suggestion is rubbished as are my views that the system won't cope in recovery. Well we are 4 days into this one and it aint over yet. Get ready for longer periods of recovery in my view (again I say it) as without a professional look at hardware in this system it can only get worse to the point of not being able to recover. Its inevitable as is the loss of the user base.

I'm doing my bit to take pressure off; I have suspended all seti work until it works properly. As I said in other threads they will get from me something equivilant to what they give me. If I get reasonable availability and performance I will give them my CPU time. If they fail I will move on a proportional basis to other projects - this I have done over the last 6-8 weeks. Today Seti has 0% from me to help them out and I suggest you all do the same to get the pressure off. I will think about how much they get back based upon their performance.

Now all ye high priests hit me again and again with your vested interest accounts of why ye wise all knowing folk always know best and anyone who complains around here is a rogue and an outcast. BS detector turned on! Flame proof suit deployed!

---

UGHH...

Now I remember why I got burned out on reading the S@H forums. Nothing but a bunch of bickering and complaining.

You are all a bunch of biddies staring at the computer screen looking for something to go wrong. When something doesn't work like it usually does you start jumping up and down and screaming about how incompetent the staff at Berkeley is.

I think I will stay with the Einstein and LHC forums. Although they move a lot slower, those people are much more positive and friendly.

Jim

---

UGHH...

Now I remember why I got burned out on reading the S@H forums. Nothing but a bunch of bickering and complaining.

You are all a bunch of biddies staring at the computer screen looking for something to go wrong. When something doesn't work like it usually does you start jumping up and down and screaming about how incompetent the staff at Berkeley is.

I think I will stay with the Einstein and LHC forums. Although they move a lot slower, those people are much more positive and friendly.

Jim

Hear, hear. Although the other projects suffer from this too, cpdn in recent weeks comes to mind, but here at least we have the ignore feature.

---

...Now all ye high priests hit me again and again with your vested interest accounts of why ye wise all knowing folk always know best and anyone who complains around here is a rogue and an outcast. BS detector turned on! Flame proof suit deployed!

Well, simply ignore the high priests (i call them cheerleaders) and let them pray their BS like a prayer mill. Beat them with their own weapons and simply ignore what they say, cause it's plain stupid simple BS-Propaganda. I've already been through this. The high priests aka cheerleaders are absolutely advisory resistant, so save your energy for something more useful ;)
Keep crunching and hope the 'people in charge' get things straight before the machinery falls apart completely *ROTFLMBAO*

---

Aloha, Uli

Keep crunching and hope the 'people in charge' get things straight before the machinery falls apart completely *ROTFLMBAO*

To the sounds of Pink Floyd

"Welcome to the Machine!"

Jim

---