Was there something wrong or have I broken something?
Result ID Claimed credit
35057894 19.24
35066517 19.24
35073206 19.24
35078819 19.24
35084448 19.24
35090108 19.24
35095409 19.24
35101421 19.24
All issued on 24-June-2006
they are still in my pending list, but they show that they have only been issued to me.
Was wondering what the issue might be.
William.Eiland
Copyright © 2024 Einstein@Home. All rights reserved.
I have 8 work units that have only been issued to me:
)
Bernd says that the scheduler had been changed to have a tendency to send shorter WUs to slower machines, in here.
Maybe this is some emergent behaviour resulting from that?
Dead men don't get the baby washed. HTH
RE: All issued on
)
Hi Bill,
There's not likely to be an issue at all - it's just the way the scheduler works. Most of the time, work is issued promptly to both members of any given quorum. Occasionally, the scheduler seems to "forget" to send out the second copy for a while. In your case it has been a week and that is a bit unusual but not unheard of. If the situation doesn't change during the next week, then raise the matter again as that might indicate a real problem.
When I looked through your results list, I noticed that you only have one result on hand on each computer, indicating that you have a nice low "connect to network" setting. Whilst EAH is normally a very reliable project, there are occasional outages like the recent aircon failure which lasted a couple of days. If you wanted to protect yourself against running out of work in such situations, you might like to increase the value of this setting to say 1.0 to 2.0 days to keep a couple of extra results on hand.
PS: Did you eventually get your Seti account sorted out?
Cheers,
Gary.
I have a similar problem with
)
I have a similar problem with one of my computers (474711). 12 results that this computer has calculated haven't yet been sent to anyone else. They do though have a second entry with a server state of "Unsent" and a red background. An example of a work unit that demonstrates this problem is 10196129 which was sent to my computer on 25 June as result 35149772. This is about the same time as I switched to using an optimised application, so the problem might be related to that. However it continued after I switched back to the standard application. On 27 June I stopped running EAH on this computer while I waited to see if the problem would rectify itself, but it hasn't done this so far.
Martin
RE: I have a similar
)
Hi Martin,
All I can do is reiterate what I said to Bill. It's not a problem at all. It's simply the way the scheduler works sometimes. In fact, if you go back and have a look at the results that Bill was talking about you will see that they have now been issued to the second computer. This happened the day after he made the initial query. Also this behaviour is totally unrelated to the particular app you were using and there is absolutely no need to stop using EAH on your computer as it will have no effect on when the scheduler will send the same work out to the second computer.
I've just had a quick look at some of my machines and lo and behold, I've found one already that is showing exactly the same thing. My guess is that this behaviour happens about 5-10% of the time, ie, unusual but not all that uncommon.
Cheers,
Gary.
Thanks Gary, I did
)
Thanks Gary,
I did get Seti worked out but have it on suspend for the time being. I have noticed that the work units in question have vanished from one of my lists but still show on the pending list. there also seems like your scheduler likes to match a fast computer with a slow on ... is that an un documented feature to keep some from running away with the numbers :-)
Bill Eiland
Gary Thanks for your
)
Gary
Thanks for your comments. About half of my unsent work units have now been sent to another computer, so it looks like they are clearing themselves in the way that you indicated. I have therefore started accepting new work on this computer.
It was the fact that the unsent work units had a red background that concerned me. Normally this indicates some sort of error. Also, this was the first time that I had seen this situation in all the time that I have been running EAH.
Martin