Status 907 Fold Log Parse Error/unknown Handle

[struan]

STATUS 907 FOLD LOG PARSE ERROR/UNKNOWN HANDLE

This has only been happening on this one computer at home. My machine(s) at work have been folding away fine, all through the ISP problems last night.

This machine is running the BeOS client, dated 13 October 2002. It is a dual-proc machine (Abit BP6), so I have two DF folders & clients running in different Terminals.

About the time of the connectivity problems, these clients were unable to upload. They were able to convert to the 129 protein, however--they now say calculating residue xxx/129. They haven't recovered today, but I wasn't getting the 905 error, either. Any suggestions???

:help: /.j

[struan]

the first
========================[ Oct 15, 2002 10:18 PM ]========================
ERROR: [000.000] {foldtrajlite.c, line 4734} Error during upload: STATUS 905 USER HANDLE NOT FOUND

my latest try:
========================[ Oct 17, 2002 9:38 PM ]========================
ERROR: [000.000] {ncbi_socket.c, line 1822} [SOCK_gethostname] Cannot get local hostname {errno=3}
ERROR: [000.000] {ncbi_socket.c, line 1822} [SOCK_gethostname] Cannot get local hostname {errno=3}
ERROR: [000.000] {foldtrajlite.c, line 5664} Error during upload: STATUS 907 FOLD LOG PARSE ERROR/UNKNOWN HANDLE
ERROR: [000.000] {ncbi_socket.c, line 1822} [SOCK_gethostname] Cannot get local hostname {errno=3}
ERROR: [000.000] {ncbi_socket.c, line 1822} [SOCK_gethostname] Cannot get local hostname {errno=3}

Be switching this machine back to seti for the time being :(

[Brian the Fist]

Please see the news page on the main web site

[struan]

10/17/2002

- We have discovered one new, strange bug that seems to occur only in BeOS, which results in a error 907 - log parse error. We are not sure what is causing the actual problem, but have fixed it anyways, so if you are using BeOS, please re-download a corrected version from the Download page. You will need to delete filelist.txt or install in a clean directory to get it working again. If this error should occur again the next time a new version is release, please let us know as we do not normally test the BeOS version.

Brian the Fist: I'm getting the same 907 error messages again with the update from yesterday. Only occurs on my BeOS machine--OSX & Windows clients updated without a hitch and are chugging away at full speed.

[Brian the Fist]

Hmm, I feared as much. I might have to 'undo' what I did last time. Let me take a look and if I find out you're right I'll post a fixed version on the web site.

[Chinasaur]

Howard,

Any chance you could quash the clients failure to kill socket threads on same bug fix?

Thanks.

[bwkaz]

Failure to kill socket threads? Is that BeOS only, or might it affect te Linux client as well? What would be the symptoms if it did?

[Chinasaur]

Yikes...my mistake. My apologies.

The socket thread issue isn't a DF problem...I posted in wrong forum. :(

Sorry for creating confusion.

[bwkaz]

Err... oh. Heh heh. Never mind then. ;)

[struan]

well, getting 907 errors on my BeOS client again after the latest update of today. windows & osX clients continue to chug away without intervention. :(

[Brian the Fist]

Ok, Im getting tired of this stupid BeOS bug. I'll put a kludge in the upload server so it will accept the malformed log files BeOS is producing as I still have no clue why it is actually doing this.
(If you recall, the problem is simply that it is eating a newline character from the log file, and the server parsing is very strict about what is on each line).
Thanks for the heads up, I'll have it accepting your data shortly so no need to re-download.

[struan]

No worries, I was surprised to see the bug again in the latest update.

I've only a single BeOS machine running df, the rest are win2k & osx boxen. I'll probably be running seti or seventeen on the BeOS machine, just to avoid running into this bug anymore. That box wasn't contributing as many df work units per unit time anyway. Time better spent running 2 seti instances, or 1 seti & a seventeen.

[Brian the Fist]

I repeat, that bug has been 'kludged' on that server. It should not ever come up again, although it is not fixed as properly as I might like.

[runestar]

Some universal bonding strips and universal bonding material (aka duct tape and bubble gum), eh, Howard? ;)

RS½