Status 908 Incorrect Protein

[ColinT]

Are we changing proteins now? Howard, what instructions do you have for us?

[Ody]

should the auto-update work this time? I have it installed as a service on my computers, so will they automatically switch over, or do I need to stop and re-start the service to get them to update?

[mnx]

just restart the client and let it autoupdate... i've got mine working on the new protien (and I'm currently 2nd for the best structure...)

mnx

[ColinT]

I hit Q and some uploaded but most of my clients simply stopped. Restarting them got the new client downloaded. This changeover abandoned thousands of WUs on my farm.

Win2K and 98se

[DATA]

had all the latest text clients on 54 cpu and every one stopped -awaiting authentication - some auto update system

why do you keep adding problems all the time - its going to take hours to get round this lot - to say i was displeased woud tell a lie

tell me/us once and for ever is it your intention to do this to every protein update

Tom
S-MDC_TRUST

[IronBits]

I had no problems, once I stopped and started the client Howard, good job.

I am using the TEXT clients only. None are running as a service on XP or win2k, and I have some running on win98se and Mandrake 8.1.

Apparently, any running clients that didn't update, were spinning their wheels crunching with the OLD data. It is not catching the "Status 908 Incorrect Protein" errorlevel, so goes back to work on the OLD cached protein again, creating more useless cached OLD WUs.

Once I stopped and re-started the client, the client looked for and found that there was a new update. It then proceeded to download the update, extracted and update itself correctly, then began work on the new protein.

However, it also leaves the OLD useless cached WUs lying about.

Getting better! :)

[DATA]

Originally posted by DATA
had all the latest text clients on 54 cpu and every one stopped -awaiting authentication - some auto update system

why do you keep adding problems all the time - its going to take hours to get round this lot - to say i was displeased woud tell a lie

tell me/us once and for ever is it your intention to do this to every protein update

Tom
S-MDC_TRUST

Just 20 more systems to do - autoupdate would have worked perfectly if it was no for the client authentication :(

Tom
S-MDC_TRUST

[Brian the Fist]

To answer a few questions at once,
Yes we have switched to protein #4 now (details in the News section of the web site).
Looks like auto-update is working properly now for most people. Your software should update itself the next time you run it or the next time it tries to upload.
Specifically to DATA with the large # of computers, what you need is to (as outlined in the README) put an 'autoupdate.cfg' file in the same directory as the client. It should contain a single line with the number 1 on it. You can create it with notepad. Alternatively, just hit 'y' to download the updates on your machines. The default requires you to hit 'y' because our stance on security is such that we don't want our software downloading updates onto your machine without your explicit approval, either by hitting 'y' or by creating the autoupdate.cfg.
Lastly, if running as a service, you do NOT need the autoupdate.cfg (but it wont hurt) and it will auto-update itself and re-start the next time it tries to upload. It's really cool,:p it really does restart the service and everything.

Any comments from our screensaver-using friends or is everyone here hardcore and using the text client?

[DATA]

Originally posted by Brian the Fist
To answer a few questions at once,
Yes we have switched to protein #4 now (details in the News section of the web site).
Looks like auto-update is working properly now for most people. Your software should update itself the next time you run it or the next time it tries to upload.
Specifically to DATA with the large # of computers, what you need is to (as outlined in the README) put an 'autoupdate.cfg' file in the same directory as the client. It should contain a single line with the number 1 on it. You can create it with notepad. Alternatively, just hit 'y' to download the updates on your machines. The default requires you to hit 'y' because our stance on security is such that we don't want our software downloading updates onto your machine without your explicit approval, either by hitting 'y' or by creating the autoupdate.cfg.
Lastly, if running as a service, you do NOT need the autoupdate.cfg (but it wont hurt) and it will auto-update itself and re-start the next time it tries to upload. It's really cool,:p it really does restart the service and everything.

Any comments from our screensaver-using friends or is everyone here hardcore and using the text client?

Thanks Howard - did this last time - but i assumed as there was so much song and dance about it not then working that this would have been built in to the client - good job i was here as should have been away - and would have ment all the systems lying ideal for 3 days and this would have been drastic

so waht is going to happen with the next protein - will it halt again - should i now add the autoupdate.cfg

Tom
S-MDC_TRUST

[Dyyryath]

Bugger!

My Windows XP laptop didn't update. I stopped it, restarted it, it asked if I wanted to update, I hit 'Y', it attempted to download, then closed. I restarted it, it asked if I wanted to update...obviously a no-go. This client had the autoupdate.cfg file. :(

My Linux workstation didn't update. It's still running. I took a peek at the directory and I'm starting to accumulate unsent units. I stopped the client, restarted it, and it automatically updated. Unfortunately, this doesn't help much as I still have to go from box to box stopping and restarting the client. This client also had the autoupdate.cfg file. :(

I know you guys are trying, but to call this a hassle for those of us with more than a couple of systems would be an understatement. :(

So far, it doesn't appear that *any* of my systems are updating without some kind of intervention on my part. :(

All of my systems have autoupdate.cfg setup. I'm launching the Linux ones like this:

nohup ./foldtrajlite -f protein -n native -df -qt > /dev/null &

I'm launching the Windows ones likes this:

.\foldtrajlite -f protein -n native -df

Each of those commands are the ones in the foldit shell script and foldit.bat batch file, respectively.

[LBaker]

I only had joy with 1 out of 5 machines. All are running win 98 or me. I had all clients updated with the latest patch and the autoupdate.cfg file was present. I still needed to stop and start the client for it to update correctly. :(

[pelligrini]

My win98 machine updated correctly, only after I restarted it again. My win2k machine won't update at all. It is using a proxy while the other isn't. There's two more win2k machines here at the office that I can't check (there are people actually working on them :rolleyes: ) Looking at all of the bz2 files on the other two I suspect they didn't update.

This morning, just to be sure, I installed the autoupdate-update from the website on all of them, they also have the autoupdate.cfg file too. :confused:

I shut down all of my clients at home before I left for the office, only leaving a win98 and a win2k machine going on DF. From what I've seen here at the office I am glad I did.

[SpongeBob SquarePants]

Ni!

I am not able to get any of my DF clients to work. I have restarted, added autoupdate.cfg, run service. I even downloaded a new client clean install. Still crunching 58 aa. Any ideas?

How many AA's is the new protein so I can check?

A very unhappy and unproductive Sponge,


SpongeBob SquarePants

[FoBoT]

Originally posted by SpongeBob SquarePants

How many AA's is the new protein so I can check?


the new one is 36

try deleting everything in the distribfold directory, then unzip it (the newest client) , then run it

[DATA]

Hi ppl

still 6 more systems to bring on line - so far not one system autoupdated (lack of autoupdate.cfg was advised) forgetting this just advised y when it requested if update was wnated and not one has refused to update all running OK - am now going to add this autoupdate.cfg file to all systems and hopefully next time round will work - but after reading above am not holding my breath - this could be painful

all systems included 6 dual systems running OSs Win98SE/Win2K Pro/Win2K Srvr and Win XP

Tom
S_MDC_TRUST

[Jammy]

Updated here too! I just closed my clients and reopened them. BTW >>> do I lose the work that was running before I updated?

Jammy



Originally posted by mnx
just restart the client and let it autoupdate... i've got mine working on the new protien (and I'm currently 2nd for the best structure...)

mnx

[pelligrini]

I got home and much to my suprise, both clients that I left running were on the new 36 sequence. The autoupdate worked on each, a win98 and a w2k. :D

[Terminator]

Just got back after a night out and all 3 home PC's have auto-updated OK and have been crunching away - Great Job Howard :)

[Angus]

None of mine - mix of NT4 and Win98 using the text client - updated automatically.

All continued to crunch the old protein, and went thru at least a couple of failed upload attempts. All have the autoupdate.cfg file in place.

Manually stopping and restarting the client caused the auto-update to work.

So- it looks like not all the bugs are worked out yet.

[Brian the Fist]

Please note that if you haven't got the latest version (and I know you're all tired of getting latest versions ;) ), some auto-update bugs were fixed in it. The latest version was released Mar. 26 (last Tuesday), whereas the previous 58 residue protein was started on Saturday the 23rd.

We have not had any problems with any of our machines running the Mar. 26 version auto-updating, except for those running behind a squid proxy server. This is because the proxy caches the status file which changes when an update occurs. If you are using a caching proxy you may experience this problem as well and should set your proxy to not cache files named 'server.status'.

Other than this I am fairly confident that the Mar. 26 version auto-updates correctly. If you know otherwise, please post a report in the Bug Report forum here and state your OS and a description of what went wrong. As long as I can reproduce the error I can fix it. It is simply not possible for us to test all ten+ versions of the software, nor can we possibly try it on the six or seven flavours of Windows in all possible modes. We are counting on your input to identify any weird and wonderful bugs that may still be remaining in the auto-update.

Thanks.

[Jammy]

My XP 1900 (NT 5.0) updated on its own with no problems. So did my T-Bird (Win 98 SE) but the 750 Classic Athlon had the yellow text which "asked" me, Y or No, if I wanted to upgarde (which I did of couse, lol).

But, I did stop the client first so I do not know whether or not they would have automatically done so themselves.

Hmm . . .I did not consider this as a bug though . . .

Jammy



Originally posted by Angus
None of mine - mix of NT4 and Win98 using the text client - updated automatically.

All continued to crunch the old protein, and went thru at least a couple of failed upload attempts. All have the autoupdate.cfg file in place.

Manually stopping and restarting the client caused the auto-update to work.

So- it looks like not all the bugs are worked out yet.

[Jammy]

Howard:

You never answered my question up above as to whether or not we will lose what work we have completed before the update became available. I noticed that when I started changing my clients that their folders had those BZ2 files in them.

Jammy



Originally posted by Brian the Fist




Other than this I am fairly confident that the Mar. 26 version auto-updates correctly. If you know otherwise, please post a report in the Bug Report forum here and state your OS and a description of what went wrong. As long as I can reproduce the error I can fix it. It is simply not possible for us to test all ten+ versions of the software, nor can we possibly try it on the six or seven flavours of Windows in all possible modes. We are counting on your input to identify any weird and wonderful bugs that may still be remaining in the auto-update.

Thanks.

[FoBoT]

the work is controlled at the server, once he makes the "change" , the server no longer accepts the work, those bz2 files are orphaned on your hard drive

it is important in the current model to pay attention to the overall project status , and upload/clear your work just before the goal (in this case 1 billion) is reached, anybody that doesn't get their work uploaded before the "Switch" at the server end is SOL

just stop/start/stop your client with "Q" and then delete and remaining .log.bz2 and .val.bz2 files, they are worthless

then restart your client and crunch away :)

[Brian the Fist]

Jammy:

As FoBoT has just said, any work left un-sent when we switch proteins is effectively lost. It would be nice if we could get that extra little bit but this is simply not possible with the manner in which our hardware and server software have been implemented (without going into boring technical details..)
If you have machines not on the network, just try to keep an eye when we get close to a full thermometer and do an 'upload dump'.
:D

[FoBoT]

hehe, its important to be "Regular" and "dump" on time :o