The Tuesday Thing?

[PinHead]

Is that an auto-update or a manual update?
Inquiring minds want to know.

Oh, is this a protein change also. Seems the month has passed?

[Brian the Fist]

autoupdate. We will be switching algorithm slightly instead of the actual protein.

[Grumpy]

So we will not need to upload our Buffered Files within 48 hours or lose them ?

[pointwood]

Ohh...Changes to the algorithm - you already found some things that could be made better in regards to the current one?

[Welnic]

Originally posted by Grumpy
So we will not need to upload our Buffered Files within 48 hours or lose them ?

I'm just guessing, but I would have to think that the 48 hour limit would apply. With a new algorithm you would want to have a new separate data set just as if it was a new protein.

[Darkness Productions]

Any word on what time this should start?

[Pascal]

Originally posted by Darkness Productions
Any word on what time this should start?

From the news section:

07/18/2003

- We will be releasing a software update on Tuesday, July 22, 2:00 PM EST. This fixes some of the bugs that have been reported.

EST = Eastern Standard Time

;)

[Pascal]

The servers seem to be down:



Server temporarily down for maintenance. Thanks for your patience.

[spldart]

I just rebooted my dually for other things and both my DF clients are not running. I started them in troubleshoot mode with dfgui and both give a message about server down. When I press enter the dos box closes and I find the client is not taking cpu cycles in task manager and is only 2.7 meg in size...
Is this related to the update?
Both copies say the same thing.

Found this in foldit.bat

:error
echo Foldtraj Client update was unsuccessful, please correct any problems,
echo ensure the file distribfold-update.exe is in the current directory and try again
pause
goto end

:end

[djp]

I just rebooted my dually for other things and both my DF clients are not running. I started them in troubleshoot mode with dfgui and both give a message about server down. When I press enter the dos box closes and I find the client is not taking cpu cycles in task manager and is only 2.7 meg in size...
Is this related to the update?
I think it is. I've got several machines hooked to the Internet running the text-based client and they each respond with an error:

Checking for newer versions...
SERVER DOWN FOR MAINTENANCE, TRY AGAIN LATER

Please try again later. Press Enter.

It looks like NOW is the time.

[pfb]

It seems if you start the client and it can't connect to the server due to maintenance, it won't start at all - unless you run it off-line...

It's a bit of a pain (only noticed as I had to reboot a PC which runs DF as a service - as soon as it couldn't connect it didn't bother to run as normal)...

[Pascal]

..
Is this related to the update?
Both copies say the same thing.

..

Yes, it is. If look into foldit.bat, you may find a line starting with

.\foldtrajlite -f protein -n native

The might be any parameters like

-uT for uplaod
-iT for internet connection
-g xx (xx is a number) for update of progress.txt

The first two -u and -i have the switches T and F for true and false. Upper or lower case is equal. You may set -iF for continuing a while until the update is done.

[tpdooley]

but he'll still have to reboot the machine to get the client running as services again; once the servers are back up. (they're controlled by service.cfg).

[Jammy]

I moved my XP 1900 and XP 2600 to Seti until the df change is completed. My T-Bird is still running df but it now has two generations buffered. It will be finishing up another generation shortly . . .does "buffered" means that the work it is doing now will still be uploaded later after the server(s) are back up? The T-Bird has been up for over six days now without a reboot.

Jammy


Originally posted by pfb
It seems if you start the client and it can't connect to the server due to maintenance, it won't start at all - unless you run it off-line...

[Jammy]

Hmm . . .if I run df as a service does that mean that I cannot use dfGUI 3.1 to visually see the work being done?


And how do I run df as a service? :confused: I use W2K.


Jammy



Originally posted by tpdooley
but he'll still have to reboot the machine to get the client running as services again; once the servers are back up. (they're controlled by service.cfg).

[Digital Parasite]

Jammy, if you read the README file that comes with the DF client, it will explain to you how to install the DF client as a Service (it works the same way on NT4, Win2k, and XP).

Yes, dfGUI will still be able to monitor your DF client if it is running as a service. The only thing different that you have to do is in the Config tab, you click on the box that says "Client Installed as Service"

Jeff.

[djp]

I just tried to upload some results and instead of getting the "SERVER DOWN FOR MAINTENANCE, TRY AGAIN LATER" message, I got a patch download and unpack. That looks good, but I didn't get the client to restart successfully. This has been on my screen for the last five minutes:

Digital signature was verified, ready to install
New version found; attempting to unpack patch file
Update was successful, restarting client
After seeing that message, I launched another client to get it to update, but this one gives me the "SERVER DOWN FOR MAINTENANCE, TRY AGAIN LATER" message.

Perhaps the server is overloaded by concurrent update requests. I think I'll let the client sit there for another half hour before I start yet another session.

[pfb]

I have 4 CPUs (3 PCs) running DF - 2 Windows and 1 Linux...the Windows PCs have downloaded the update but crash after doing one pass of a 96 protein - even though the update email I got said it would be 127...

anyone else got the 'updated' client and experiencing problems? I am going to take all the DF clients off-line for the moment...

[Welnic]

If you need to start the client while the ""Server is down for maintenance" you have two options.

Edit your foldit.bat file if you are running the command line client and add -if. If you are running as a service then you need to add connect=0 to the service.cfp file. Then once it comes back up you can either delete what you added or change them to -it or connect=1. This makes it easier to remember what to do next time.

The other option is to unplug your ethernet cable and start the client. It will take a little while to realize that it can't find the server and then it will start crunching. Now plug the cable back in. When it tries to upload if it sees that the server is down it will keep folding and try again later. When the server comes back it will just proceed normally.

[nickbrownsfan]

Same problems as everyone else. Windows XP pro client dead said update loaded unpacked and restarting client but nothing else happening. All the other machines couldnt connect. Will run off line for a bit or switch to another project until server are not so swamped.

[RipItUp]

Originally posted by pfb
I have 4 CPUs (3 PCs) running DF - 2 Windows and 1 Linux...the Windows PCs have downloaded the update but crash after doing one pass of a 96 protein - even though the update email I got said it would be 127...

anyone else got the 'updated' client and experiencing problems? I am going to take all the DF clients off-line for the moment...

I got this on two machines,

gave an illegal operation performed.

My other machine hung as per the email before after trying to restart the client.

The other machines cannot connect at all at present, if they could I guess I will have all 5 machines getting illegal operatrions ....

Regards

Andy

[nickbrownsfan]

Ok the windows xp pro machine after attempting to run it offline is also getting the ill. op. error and shuts itself off. :( The others which I dont think grabbed the update are running off-line but that is with win98.

[spldart]

Thanks for the -if guys...My machines are running again 'foldtrajlite' pulling all extra cpu usage...But some of the incidences on some machines aren't using their normal amount of memory :confused: All are running with -rt and -qt as well.
I hope I get credit for work done on the old client while they are sorting through this server thing.

[Mikus]

I'm trying to load from the (old) client every 30 minutes, but I always get the SERVER DOWN FOR MAINTENANCE message (after "checking for new version"). That doesn't surprise me: __NONE__ of the autoupdates over the past many many months have worked for me (before that, they *did* work).

What upsets me, though, is that I kept getting a screen saying "Server temporarily down for maintenance" whenever I tried to access http://www.distributedfolding.org. [Finally got through to the site at 1700 CDT]

I don't mind having to do a manual download (been doing it for every client/protein change for these many months now). But WHY must I wait due to WWW server "unavailability" ?

I'd much prefer if the new client were available for WWW upload 24 hours (or more) *before* the scheduled switchover time (the DF data server can enforce not recognizing new-client fileset uploads before the proper time). Then I would get the new-client download over with ahead of time, and would perform the client switchover when I got the "Server Down" message fom the old client. I don't want to wait HOURS after the scheduled switchover time for the WWW server to become available.

mikus (using Linux)

[Jammy]

Hey Mikus . . .I hear ya and wholeheartedly agree with you!

Jammy




Originally posted by Mikus


I don't mind having to do a manual download (been doing it for every client/protein change for these many months now). But WHY must I wait due to WWW server "unavailability" ?

I'd much prefer if the new client were available for WWW upload 24 hours (or more) *before* the scheduled switchover time (the DF data server can enforce not recognizing new-client fileset uploads before the proper time). Then I would get the new-client download over with ahead of time, and would perform the client switchover when I got the "Server Down" message fom the old client. I don't want to wait HOURS after the scheduled switchover time for the WWW server to become available.
[/B]

[Mikus]

Latest news:

At 1730 CDT the "check for new version" __did__ download the latest client "automatically" -- the first time in many many months. [I don't have the authorization token in my directory, so it asked me, and I said yes.]

Though I had started with flags -rt -u t -- the download of the new client auto-continued with my regular 'foldit' (flags -rt -g 1 -i f), starting with a new generation_0. I issued 'Q' and restarted with flags -rt -u t -- and it *did* upload the filesets accumulated since the previous upload (done at 1300 CDT - before the scheduled switchover). Then I did my "regular" startup of 'foldit', and it is now cranking away.

mikus (all's well than ends well)

[Welnic]

After the update is downloaded the regular foldit is called to start the new client, so the behavior you saw is what it is supposed to do.

[Jammy]

Well I had four buffered structures on two boxes and three on another and it seems that they were not counted although I was able to upload them before the newest client started (I had a hard time but I did it and I don't even remember how I did it!).

Anyways . . .it does not seem that those 10 generations were even counted!

Jammy




Originally posted by Welnic
I'm just guessing, but I would have to think that the 48 hour limit would apply. With a new algorithm you would want to have a new separate data set just as if it was a new protein.

[Brian the Fist]

Originally posted by Mikus
I'm trying to load from the (old) client every 30 minutes, but I always get the SERVER DOWN FOR MAINTENANCE message (after "checking for new version"). That doesn't surprise me: __NONE__ of the autoupdates over the past many many months have worked for me (before that, they *did* work).

What upsets me, though, is that I kept getting a screen saying "Server temporarily down for maintenance" whenever I tried to access http://www.distributedfolding.org. [Finally got through to the site at 1700 CDT]

I don't mind having to do a manual download (been doing it for every client/protein change for these many months now). But WHY must I wait due to WWW server "unavailability" ?

I'd much prefer if the new client were available for WWW upload 24 hours (or more) *before* the scheduled switchover time (the DF data server can enforce not recognizing new-client fileset uploads before the proper time). Then I would get the new-client download over with ahead of time, and would perform the client switchover when I got the "Server Down" message fom the old client. I don't want to wait HOURS after the scheduled switchover time for the WWW server to become available.

mikus (using Linux)

Perhaps you should try downloading the 'autoupdate proxy daemon', this does exactly what you are asking for.

[^7_of_9]

Originally posted by Brian the Fist
Perhaps you should try downloading the 'autoupdate proxy daemon', this does exactly what you are asking for.

I did. Don't like it. The thing generates so many errors that go into the the error log that I had it up and running for a week and the error log was almost 100MB!! Then I don't know if it was jsut because of this particular update or what, but my machines didn't seem to take the update from it at all even though I checked and it was downloaded ok. Maybe somehow the log can be trimmed or something somehow?