61) Message boards : Number crunching : my client will never connect (Message 18219)
Posted 16 Oct 2007 by PovAddict
Post:
Me too!! I realize you can see the current deferment in the projects tab, but it is nice to see the history (like 1 min deferrals growing to 2 hours or more, etc) so you have a better idea of what's going on.


My MSG tab shows deferral amounts and why. (.10.20)

The next version of BOINC will not show it anymore, that's what I was talking about. I think five people against the change is enough to get it reverted :)
62) Message boards : Number crunching : my client will never connect (Message 18218)
Posted 16 Oct 2007 by PovAddict
Post:
What I'm trying to point out is that LHC does not seem to be communicating the required backoff time to the client. Maybe this is because doing so requires the server code upgrade they've been mentioning.

No, the bug was client-side, v5.10.14 introduced a bug that means every time you asks for work but don't get any, the deferral-message from project are not used.

FYI: it's now fixed, fix will be available on 5.10.23.
63) Message boards : Number crunching : The new look bugs (Message 18208)
Posted 16 Oct 2007 by PovAddict
Post:
Any hope of getting jump to first unread post fixed?

What she said :)
64) Message boards : Number crunching : server 4.67 ???? (Message 18180)
Posted 16 Oct 2007 by PovAddict
Post:
I thought you were gonna update??? dont go backwards!!!!

Where did you get that version number? The client says LHC has scheduler version 5.05.
65) Message boards : Number crunching : my client will never connect (Message 18175)
Posted 15 Oct 2007 by PovAddict
Post:
(In [13849]) - client: fix bug where delay request from project is ignored if no results are returned.

Expect the fix on 5.10.23. Now for the bad part: the "Deferring communication for..." is now shown only if you have debugging enabled. Who thinks the delay should be always shown?
66) Message boards : Number crunching : Ghost result (server thinks I got it, but I didn't) (Message 18172)
Posted 15 Oct 2007 by PovAddict
Post:
According to Neasan (or was it Alex?) they are going to update it within the next couple of weeks.


a) It was according to me
b) I never gave a time scale (I'm not stupid)

It is the next priority obviously but it won't be 10 minutes of dragging and dropping files(which some seem to believe it is). We now more or less know how this whole think works now and this will make upgrading easier and de-bugging easier once it is done but it won't happen over night.

I will investigate this but assume (as POVaddict seems to love mentioning) this will involve the server upgrade.

That tag seems to exist since the scheduler has an XML parser (24/05/2006 22:22:10). The functionality itself goes back to over two years ago, although I have no idea how it was enabled back then.
67) Message boards : Number crunching : Ghost result (server thinks I got it, but I didn't) (Message 18139)
Posted 14 Oct 2007 by PovAddict
Post:
Those of us who participate with SETI@Home are familiar with this kind of thing happening. The admins at SETI figured out how to resend those resultIDs to the volunteers (or so I recall...someone correct me if I'm wrong)...

No need to "figure out" too much. It's just a setting that has existed for-freakin-ever: <resend_lost_results>.
68) Message boards : Number crunching : Please increase our daily machine quota (Message 18127)
Posted 13 Oct 2007 by PovAddict
Post:
10/13/2007 10.29.35|lhcathome|Message from server: No work sent
10/13/2007 10.29.35|lhcathome|Message from server: (reached daily quota of 4 results)


That's on a dual-core, is the limit 2/day per core?

Yes it is. If you look on your computer details it says "Maximum daily WU quota per CPU: 2/day".

EDIT: admins, why is bold text also 10% bigger?
69) Message boards : Number crunching : work units?? (Message 18126)
Posted 13 Oct 2007 by PovAddict
Post:
Yes. Today ( Oct 13, 2007 ) shows Up, 52030 workunits to crunch and when I hit the update I still do not get a download.

Whats up with this?

Read the whole thread, and have a look at your Messages tab. If you don't wait an hour between updates, you won't get anything. Don't click the update button manually.
70) Message boards : Number crunching : Claimed 0 credits - still pending after all 5 results received (Message 18120)
Posted 13 Oct 2007 by PovAddict
Post:
Interestingly strange. If the results don't match, the validate state would be "no consensus yet". If some of the results are invalid, the validate state would be "invalid". In both cases, an extra replica would have been sent. None of this happened; validate state is "initial" which I think means the validator hasn't even looked at the results yet. Server status doesn't show any big validation backlog
71) Message boards : Number crunching : Please increase our daily machine quota (Message 18117)
Posted 13 Oct 2007 by PovAddict
Post:
I would be very happy with a 2 pending wu's per core quota. So that when I finish one, I can grab another. Having one backup wu is nice in case there's a network problem or whatever.

That will be possible after they upgrade the server.
72) Message boards : Number crunching : Happenings... (Message 18110)
Posted 13 Oct 2007 by PovAddict
Post:
No problem with new members joining but I am one of the basic 12 members that have been logged on the site 24/7 for several years now even when there was many months of nothing just to stay in touch here.

It is just hard to look at the mountain of wu's and not get to at least have em as fast as I can finish them.

Once the new users are happy (not finding an inactive project), AND the server is upgraded, they can start using fair and non-time-wasting work distribution settings. For now, it's better that everybody can get some (although in my opinion 2 a day is way too low, but well).
73) Message boards : Number crunching : Please increase our daily machine quota (Message 18102)
Posted 12 Oct 2007 by PovAddict
Post:
I'm with MrWizard! You should increase again the daily quota to 4 or 8 because there are more than 80000 WU's to crunch. I now that there are many users but the faster work is made the faster you can get new work.

Did you read the reason why the quota was decreased?

Many new users are expected soon because of some recent "advertising". They better find something to do when they come.

If the quota is too large, we'll have the usual problem of some people with hundreds and some people with nothing. New users come, get no work, go away quick. If everybody, including the new users, gets 2 workunits a day, they will stay seeing the project is active.

But the batch of work will get done slower. Is it worth it? The admins expect to upgrade the server soon (RIGHT? >.>) and that would let them use actually fair work scheduling. Meanwhile, we're in "show new users we're alive" mode, not in "fair work distribution" mode.
74) Message boards : Number crunching : Happenings... (Message 18091)
Posted 12 Oct 2007 by PovAddict
Post:
BTW - It seems the quota may actually be set at 4 WUs. The machine that first started getting work received 4 before the server message said quota reached on the request for number 5. Results for computer

It's set to two WUs per CPU. That computer is dualcore, so 4 for it.
75) Message boards : Number crunching : Please increase our daily machine quota (Message 18088)
Posted 12 Oct 2007 by PovAddict
Post:
How about reading other threads before posting? It was decreased for a reason.
76) Message boards : Number crunching : Happenings... (Message 18083)
Posted 12 Oct 2007 by PovAddict
Post:
Two workunits a day, one request an hour.


So, the quota applies to CPU while the request time applies to user.
Just fate that first box received a WU before the second box asked. The second box was told NO. Then the first box requested and received work after the hour delay, but before the second box could try again.
First box happy. Second box pouting and waiting.

No, request time applies to individual machines too.

One problem I see is that the server requests a 1-hour delay and the client isn't following it. I'm not sure if doing another request before the hour makes the last-request-timer on the server "reset" (which would make you NEVER get work if you request every 30 minutes). It certainly shouldn't.
77) Message boards : Number crunching : Happenings... (Message 18072)
Posted 12 Oct 2007 by PovAddict
Post:
... the quota is per CPU not user or IP so people with more than one machine should get 2 on each.


Thank you for clarifying how the quota restrictions are applied.

However, I am still puzzled by server message traffic when requesting work. I have two machines attached to this project. One received 2 WUs without difficulty. The second is repeatedly rebuffed with the explanation:

  • Message from server: Not sending work - last request too recent: xxxx sec
  • Deferring communication for x hr x min xx sec



Two workunits a day, one request an hour.
78) Questions and Answers : Windows : Constant "last request too recent" error (Message 18055)
Posted 12 Oct 2007 by PovAddict
Post:
Both of my machines have been getting this error constantly. BOINC backs off for a minute or two then tries again. Same error, and BOINC waits longer. Keeping repeating this...

...now my machines are waiting around 1 hour. I am not manually updating the project - BOINC is doing everything.

What's going on? Why can't I get any work units? :/

I think that's a BOINC bug. If the server tells it to wait an hour, it should wait an hour. It currently seems to be retrying in a minute and getting the "last request too recent - I TOLD YOU TO WAIT AN HOUR" message.
79) Message boards : Number crunching : Merge?? (Message 18054)
Posted 12 Oct 2007 by PovAddict
Post:
Good code: use POST instead of GET. I think this has also been fixed months ago. Admins: Are you going to upgrade the server code some time in this century?
80) Message boards : Number crunching : work units?? (Message 18053)
Posted 12 Oct 2007 by PovAddict
Post:
So that they have a fighting chance of getting a job to process we changed some of the settings including the daily quota.

2 workunits per day? I want the drugs you're taking.


Previous 20 · Next 20


©2024 CERN