New beta version bugs

[Stardragon]

Please post bugs related to the newly available beta version (http://www.free-dc.org/forum/showthread.php?s=&threadid=5803) in this thread.

In particular, I am interested to see if the random characters appearing in filelist.txt are still a problem that persists. IF you do encounter this error, please be as detailed as possible in your report as to the conditions you were running with (offline then upload, constantly online, etc) as well as ALL the flags that were being used - both before and after. This error would most likely occur when the client is exited and then restarted, so please pay close attention to those times when testing.

If you can come up with a way to recreate the error, that would be even more beneficial towards fixing it.

[RandomCritterz]

Twice I've had the new beta lock-up, requiring a re-boot. Nothing running but Win98SE and ZoneAlarm, folding using this .bat
echo "Fold!" > keepit.lock
:start
.\foldtrajlite -f protein -n native -qt -if -rt -p-15 -g10
.\foldtrajlite -f protein -n native -ut
if EXIST keepit.lock goto startwith a scheduled task to delete foldtrajlite.lock every 12 hours.

Both times, it has locked when starting to upload. Both times the end of the error.log showed the "========================[ Date Time ]========================" line but not "Starting foldtrajlite built Mar 9 2004". Upon restart, folding resumes without additional errors.

Win98SE, 512MB, 900 MHz Athlon.

[RandomCritterz]

The garbage characters only showed up that first weekend when the beta-server was borked. Perhaps you can schedule a time for it to act up again, so we can see if the newer beta can cope better.

[Stardragon]

Originally posted by RandomCritterz
Twice I've had the new beta lock-up, requiring a re-boot. Nothing running but Win98SE and ZoneAlarm, folding using this .bat
echo "Fold!" > keepit.lock
:start
.\foldtrajlite -f protein -n native -qt -if -rt -p-15 -g10
.\foldtrajlite -f protein -n native -ut
if EXIST keepit.lock goto startwith a scheduled task to delete foldtrajlite.lock every 12 hours.

Both times, it has locked when starting to upload. Both times the end of the error.log showed the "========================[ Date Time ]========================" line but not "Starting foldtrajlite built Mar 9 2004". Upon restart, folding resumes without additional errors.

Win98SE, 512MB, 900 MHz Athlon.

The first culprit I wold point out is running with such high priority (-p-15) on a Win98 machine. From your description, it is unclear to me how you even know the script reaches the upload stage. I suspect your scheduled task of deleting the lock file may not get executed, or the system locks up while attempting to remove exit the client (possibly lack of resources).

Try running the client at regular priority, as well as manually stopping it before an upload.

[Stardragon]

Just a reminder, in case I was somewhat vague before, for you to report any client errors from the most recent beta - both old errors that you may have seen before and new errors as well.

[RandomCritterz]

Originally posted by Stardragon
The first culprit I wold point out is running with such high priority (-p-15) on a Win98 machine.This box has been dedicated to folding since July, with priorities of -15 to -18, without previous lock-ups. :shrug: I'll take it down a notch, and see what happens.
From your description, it is unclear to me how you even know the script reaches the upload stage. I suspect your scheduled task of deleting the lock file may not get executed, or the system locks up while attempting to remove exit the client (possibly lack of resources).I can tell by the error.log. If foldtrajlite.lock hadn't been deleted on schedule, or the -ut hadn't started, error.log wouldn't show the appropriate start time.

I'll fire up a monitoring program to track resources - does this beta use more than the previous eight months worth of clients?

What does the client do between writing its start time and its build date?

[Galuvian]

The new version locked up on me while making trajectory distribution. Restarting the client was enough to get it crunching again.

There was nothing related to the lockup in the log file.

Running the default foldit.bat file
Win2k SP2
Dual Xeon 1.7Ghz
2GB RAM

Do you need any other information?

[erk]

I read this thread after the problem occured so I had already trashed my filelist.txt without thinking. Nextime I will pay more attention. I am running the beta client on linux kernel:

Linux athlon2 2.4.21 #3 SMP Thu Oct 9 12:07:06 CEST 2003 i686 unknown unknown GNU/Linux with 512MB of RAM on a dual athlon MSI K7D MP2800+ one CPU running the regular client as a speed comparision.

The error I got when I tried to purge the filelist was:


/foldtrajlite -purgeuploadlist 1
[foldtrajlite] FATAL ERROR: [000.000] Illegal file remark 3: Ab Initio Random Protein Structure found in upload list
Hit Return

There was no apparent reason for the problem, the machine does nothing but DF at present, on a very reliable Internet connection.

If it does it again, what additional info would you like, apart from the corrupt copy of filelist.txt?

[pointwood]

Nothing new, but I thought I would report it anyway. I get *a lot* of these in the error log:



Fri Mar 12 20:37:47 2004 ERROR: [777.000] {ncbi_socket.c, line 1258} [SOCK::s_Connect] Failed pending connect to anteaterbeta.blueprint.org:80 (Unknown) {errno=No such file or directory}
Fri Mar 12 20:37:47 2004 ERROR: [777.000] {ncbi_connutil.c, line 801} [URL_Connect] Socket connect to anteaterbeta.blueprint.org:80 failed: Unknown


Furthermore I got this in my error log:

Sat Mar 13 11:57:05 2004 ERROR: [000.000] {foldtrajlite2.c, line 4963} Error during upload: STATUS 906 GENERIC STRUCTURE ERROR
Sat Mar 13 17:34:01 2004 ERROR: [000.000] {foldtrajlite2.c, line 4963} Error during upload: STATUS 906 GENERIC STRUCTURE ERROR
I haven't seen that one before.

My client is still running though, so I'm curious about how serious this error is.

[Stardragon]

Originally posted by erk
The error I got when I tried to purge the filelist was:

There was no apparent reason for the problem, the machine does nothing but DF at present, on a very reliable Internet connection.

If it does it again, what additional info would you like, apart from the corrupt copy of filelist.txt?

Please post the exact condition of the client prior to the error - i.e. what were you doing previously, why you needed to run purgeuploadlist, etc.

FOR EXAMPLE: was folding offline, changed flags to -qf -ut to upload results, and the error occured after upload was done and client was trying to restart folding.

Also, just to double-check, are you running the second (newest) version of the beta client?

[Stardragon]

Originally posted by pointwood

Furthermore I got this in my error log:

Sat Mar 13 11:57:05 2004 ERROR: [000.000] {foldtrajlite2.c, line 4963} Error during upload: STATUS 906 GENERIC STRUCTURE ERROR
Sat Mar 13 17:34:01 2004 ERROR: [000.000] {foldtrajlite2.c, line 4963} Error during upload: STATUS 906 GENERIC STRUCTURE ERROR
I haven't seen that one before.

My client is still running though, so I'm curious about how serious this error is.

Again, can you tell me what flags you were running with, and what was the client doing when you received this error? The surrounding context of the error log before and after these 906 errors would also be beneficial.

[pointwood]

It was running hidden (not as a service) and with the mem option. It was doing nothing besides crunching at the time since it was happening this past weekend.

The only thing otherwise in the log is a lot of the other lines I posted. I'm on a 2Mbit pretty stable permanent line.