BUG 1 - "Unable to read trajectory distribution"

[Stardragon]

I'm trying to take a systematic approach to the bugs currently appearing, so could you please keep all related postings in the relevant bug thread.

First I would like to address the error that reads as follows:
[NULL_caption] FATAL ERROR: [002.003] Unable to read trajectory distribution, please create a new one

I'd ask users to try the following 2 things:
1) Stop and close the client, then rename the *.trj file to the previous generation. So if you have "yourhandle_protein_15.trj", rename it to "yourhandle_protein_14.trj"
Restart the client and see if things are running ok - the results should be uploading.
If this indeed fixes the error, please post to this thread telling me it worked.

2) If the above suggestion does not work for you, please post here letting me know the following:
- screensaver or client
- what flag settings you have
- with/without DFGUI
- is it a service (Win only)
- what OS you're running on

Thank you for all your feedback so far, I hope this will be fixed soon. And if I don't reply for a while, this means I'm busy bug hunting, so just me.

[YoKenny]

> 1) Stop and close the client, then rename the *.trj file to the previous generation. So if you have "yourhandle_protein_15.trj", rename it to "yourhandle_protein_14.trj"

I do not have this file in the folder.

> 2) If the above suggestion does not work for you, please post here letting me know the following:
- screensaver or client

screensaver

- what flag settings you have

Auto install Yes
Temporary No
Use up to Yes

- with/without DFGUI

?

- is it a service (Win only)

I don't think so.

- what OS you're running on

WinXP SP1 with all Windows Updates

[YoKenny]

More information.

I renamed filelist.txt to filelist.old and used the Preview in screensaver setup and it does not get the error message.

[aptarasc]

The handle_protien_gen.trj file does not exist
using screensaver

[wirthi]

Hi,

point #1 worked for me.

Client was ungently shutdown by my notebook when it ran low on battery. After restarting the notebook and the client I got the error you described above. Renaming the described file solved the problem (at least the computer is folding again now).


It didn't solve a similar problem on another client though. This client has 21 files in it's buffer. When I start it, the following lines are output:

Checking for newer versions ...
Uploading fileset 1/21 from last run to server...
Previous generation missing

Some seconds later the "Handle found. Using ....." message is displayed. After that the "Unable to read trajectory distribution"-error flollows. I cannot repair this client with the method you described.

Using the textclient with "-rt -g 5" switches (most likely it was "-qt" while the error occured for the first time). I usually use dfGUI (3.01) but as soon as I recognized there is something wrong I tried to solve it without dfGUI. NO-service (text client). OS is WinXP Home SP1, German edition.

Greets,
Christian

[small_furry_spider]

YoKenny- renaming filelist.txt will get it running again, but it will start from generation 0 again. This is not ideal as you get more points for higher generations and, more importantly, higher generations optimise the fold from the 0th generation so you will never get good energies. (If you don't know what dfGUI or service is then you aren't running them )

Stardragon- I have this problem. Renaming yourhandle_protein_15.trj to 14.trj gets it running again, but it does not look like it is uploading as I now have:

yourhandle_0_yourhandle_protein_15_0000003.val
yourhandle_0_yourhandle_protein_15_0000003_min.val

and numbers upwards (16,17,18,19 etc) in the folding directory.

However I don't have:

yourhandle_0_yourhandle_protein_15_0000003_min.val.bz2
(and 16,17,18 etc)

in my folder. I am not sure what these 3 files are. How do I tell if it is uploading properly (but not deleting files properly) or if no uploading is happening at all? I have 5 machines folding so I cannot easily tell from my totals.

[tpdooley]

During the beta testing - we noticed that the generation numbers storied on the HD were off by one; and it was fixed in the next beta. Did we get rolled back to the client prior to that fix?