One of my computers has 20 completed work units but will not upload them and retrieve new work. I see this in the log:
3/11/2014 11:18:06 AM | Einstein@Home | [error] Can't parse workunit in scheduler reply: unexpected XML tag or syntax
3/11/2014 11:18:06 AM | Einstein@Home | [error] No close tag in scheduler reply
Any idea on correcting this?
Copyright © 2024 Einstein@Home. All rights reserved.
[error] Can't parse workunit in scheduler reply
)
You have updated 7 of your 10 computers from BOINC v7.2.39 to v7.2.42
If this is one of the ones you haven't updated yet, I suggest you do so and try again. v7.2.39 has a bug which can block communications.
This particular computer is
)
This particular computer is running v7.2.42
Mind telling which hostID it
)
Mind telling which hostID it is? Perhaps that its scheduler log shows us something.
ID: 6242068
)
ID: 6242068
OK, it shows that host asking
)
OK, it shows that host asking for work.
It shows that all the work it just reported was already reported before.
And it shows it sent that host 1 lost task
So what does it now show in the event log of that host?
All those messages sound
)
All those messages sound right for a host which has sent in a sched_request, but failed to parse the sched_reply - and hence sent the same set of requests in again.
If it doesn't clear up by itself (perhaps after restarting BOINC), we might have to ask for copies of those two files.
I ended up just uninstalling,
)
I ended up just uninstalling, removing everything in c:\program data\BOINC, and re-installing. Not sure what happened, but it's working again.
Thanks,
Well, I take that back, it's
)
Well, I take that back, it's still not working correctly. It downloaded two work units and now I'm getting this again:
[error] Can't parse workunit in scheduler reply: unexpected XML tag or syntax
[error] No close tag in scheduler reply
Sounds as if there's a lack
)
Sounds as if there's a lack of an EOL character in the sched_reply*.xml file, just before the last end-tag. But why it's on your system only, I don't know. I'll try to get one of the admins to look in here, and if need be, you'll have to send your sched_request_einstein.phys.uwm.edu.xml and sched_reply_einstein.phys.uwm.edu.xml files to him, so he can check that they're doing what they're supposed to do.
Don't post them here, please, as they contain your account key. With that anyone reading could do harm to your account.
Now, I see that your hostID changed, it's now #10728002, if I am not mistaken. So then here's a question in general, why is hostID #6242068 now all of a sudden an anonymous platform, with zero credit?
OK, an update. There's been a
)
OK, an update. There's been a flurry of activity in the background, since I emailed the administrators & moderators about this. This would appear to be a problem when the scheduler reply doesn't fit into the XML buffer of the BOINC client. This happens when there are many files required to download for a task (as it is the case for GW (S6CasA) work) and there are many download mirrors available (but only 4 in this case).
There are a couple of options to remedy this. Some through the server-side, one requires an as-off-yet-not-available updated BOINC client (so no need to look for it) and one could be done by you. The administrators prefer to try the server sided changes first. Just in case someone else runs into the same problem.
In any case, don't do anything yet. I'll post a further update later on.