Folding Forum

Been having this problem for a while. It has always resolved itself so far, but an awful lot of processing time is wasted. After uploading results, FAH tries unsuccessfully over and over to get the next work packet. It starts by saying:

[16:03:25] - Preparing to get new work unit...
[16:03:25] + Attempting to get work packet
[16:03:25] - Connecting to assignment server
[16:03:26] - Successful: assigned to (0.0.0.0).
[16:03:26] + News From Folding@Home: Welcome to Folding@Home
[16:03:26] Work Unit has an invalid address.
[16:03:26] - Error: Attempt #1 to get work failed, and no other work to do.
Waiting before retry.

Then goes on to say

[16:03:36] + Attempting to get work packet
[16:03:36] - Connecting to assignment server
[16:03:37] - Successful: assigned to (0.0.0.0).
[16:03:37] + News From Folding@Home: Welcome to Folding@Home
[16:03:37] Work Unit has an invalid address.
[16:03:37] - Error: Attempt #2 to get work failed, and no other work to do.
Waiting before retry.

And repeats the pattern over and over. Every attempt says 'successful' but then says 'invalid address'

Eventually I get a message like

[23:18:17] + Attempting to get work packet
[23:18:17] - Connecting to assignment server
[23:18:18] - Successful: assigned to (171.64.122.83).
[23:18:18] + News From Folding@Home: Welcome to Folding@Home
[23:18:18] Loaded queue successfully.
[23:22:38] + Closed connections
[23:22:38]
[23:22:38] + Processing work unit

On this occasion it took seven hours of attempts before I got something workable off the assignment server. So.
How come I keep getting (0.0.0.0) which is invalid?
And if I can get a valid assignment after multiple attempts without reconfiguring FAH, what's stopping me getting one right away?
Is there a short cut I can use when I get these hang ups?
Using XP Home SP2 on a Broadband connection.

Thanks

Hello John Dean, welcome to the forum.

That could be a network problem between you and Stanford, at Stanford, or a SMP work server out of work units or even offline. It usually comes back later. All part of a beta client.

See also: http://fahwiki.net/index.php?title=Why_ ... 0.0.0.0%3F

John Dean wrote:And if I can get a valid assignment after multiple attempts without reconfiguring FAH, what's stopping me getting one right away?
Is there a short cut I can use when I get these hang ups?
Using XP Home SP2 on a Broadband connection.

You didn't specify which FAH client you were using nor which configuration settings were involved and both can contribute to this sort of problem. The Assignment Server has an extremely complex job of picking the right server that will have WUs that match your client and settings and it gets a lot tougher when some of the servers are short of WUs.

There are no shortcuts.

All WUs (except Gen=0) are created from returned WUs. Thus a server that was out of work a few minutes ago might have new WUs to assign a few minutes later.

bruce wrote:

John Dean wrote:And if I can get a valid assignment after multiple attempts without reconfiguring FAH, what's stopping me getting one right away?
Is there a short cut I can use when I get these hang ups?
Using XP Home SP2 on a Broadband connection.

You didn't specify which FAH client you were using nor which configuration settings were involved and both can contribute to this sort of problem.

I'm using FAH for Windows with XP Home. I don't know which configuration settings are of interest - which ones do you need to see?
Following the link that Zim gave, I see "The Assignment Server now goes through a logical set of conditions and if it can't find a match, it assigns to 0.0.0.0, which means the client will just try again later. If this continues for an extended period of time (more than 4 hours), be sure to ask for help in the Folding-Community support forum because this problem is normally corrected quickly."
This suggests there is something that can be done but I often wait for much longer than 4 hours. In the latest case, 7 hours. This is a fairly recent problem for me. (Despite the indicator that this is my first posting here, I've been around for a couple of years but had to re-register in the Great Server Blow-Up.)
Zim's reference also quotes "The SMP client requires the >5MB (Big WU) option set to Yes. A No setting will often result in being assigned to 0.0.0.0. Setting the SMP client (and possibly the GPU and CPU clients) to request deadlineless WUs may also cause the 0.0.0.0 work server assignment."
I have FAH configured to receive large WUs and I have "no preference" set for deadlines

Changing or removing client switches can sometimes help. +/- advmethods, or any other custom switches, and try again.

Can you ping any Stanford servers, like 171.64.65.64 or 171.64.65.56?

The client can connect to the servers on two ports. Port 8080 is the preferred method and it's always tried first ( http://assign.stanford.edu:8080 ). Some proxies block connections to an external port 8080, though, so there is a second alternative. If that connection cannot be opened, the fall-back condition is to connect to http://assign2.stanford.edu port 80. [Your log looks like you failed on port 8080 and are using port 80.] Once you connect to either assignment server, you'll be redirected to a Work Server. Fewer Work Servers are available on port 80, so you're more likely to get assigned to 0.0.0.0 when the AS finds there are no WS's that match your settings. The situation is dynamic, so you may see a reasonable number of assignment failures before you get an assignment that works. There also may be something that the Pande Group needs to correct (such as adding more WUs to a port 80 server). If it has been 4 hours (I'd probably reduce that number, but 7im isn't wrong when he suggests it) somebody from the Pande Group needs to know about it. (They may have already noticed or we may need to notify them.) How long it takes to fix it after that happens depends on what they need to do to fix it.

bruce wrote:If it has been 4 hours (I'd probably reduce that number, but 7im isn't wrong when he suggests it) somebody from the Pande Group needs to know about it. (They may have already noticed or we may need to notify them.) How long it takes to fix it after that happens depends on what they need to do to fix it.

It's usually at least four hours for me and, as I said, the last one was 7 hours. So how would I notify Pande Group?
NB The current unit loaded comparatively quickly:

[23:05:17] - Preparing to get new work unit...
[23:05:17] + Attempting to get work packet
[23:05:17] - Connecting to assignment server
[23:05:18] - Successful: assigned to (171.65.103.160).
[23:05:18] + News From Folding@Home: Welcome to Folding@Home
[23:05:18] Loaded queue successfully.
[23:05:19] + Could not connect to Work Server
[23:05:19] - Error: Attempt #1 to get work failed, and no other work to do.
Waiting before retry.
[23:05:29] + Attempting to get work packet
[23:05:29] - Connecting to assignment server
[23:05:30] - Successful: assigned to (171.65.103.160).
[23:05:30] + News From Folding@Home: Welcome to Folding@Home
[23:05:30] Loaded queue successfully.
[23:05:31] + Could not connect to Work Server
[23:05:31] - Error: Attempt #2 to get work failed, and no other work to do.
Waiting before retry.
[23:05:47] + Attempting to get work packet
[23:05:47] - Connecting to assignment server
[23:05:48] - Successful: assigned to (171.64.122.128).
[23:05:48] + News From Folding@Home: Welcome to Folding@Home
[23:05:48] Loaded queue successfully.
[23:35:49] - Couldn't send HTTP request to server
[23:35:49] + Could not connect to Work Server
[23:35:49] - Error: Attempt #3 to get work failed, and no other work to do.
Waiting before retry.
[23:36:13] + Attempting to get work packet
[23:36:13] - Connecting to assignment server
[23:36:14] - Successful: assigned to (171.64.65.102).
[23:36:14] + News From Folding@Home: Welcome to Folding@Home
[23:36:14] Loaded queue successfully.
[23:36:25] + Closed connections
[23:36:25]
[23:36:25] + Processing work unit
[23:36:25] Core required: FahCore_78.exe
[23:36:25] Core found.
[23:36:25] Working on Unit 04 [January 25 23:36:25]
[23:36:25] + Working ...

I seem to have touched all the bases with the servers and only waited half an hour.

John Dean wrote: So how would I notify Pande Group?

Posting here is usually sufficient.

One comment though... we haven't had a lot of others posting the same problem yet, so the symptom seems local to your network, client settings or client type, etc. If Stanford had been down for 7 hours, (other than the scheduled power outage posted on the News page) then we would have lots of posts about it.

7im wrote:One comment though... we haven't had a lot of others posting the same problem yet, so the symptom seems local to your network, client settings or client type, etc. If Stanford had been down for 7 hours, (other than the scheduled power outage posted on the News page) then we would have lots of posts about it.

Sure.
But my log entries didn't suggest the servers were down. A typical entry was:

[16:03:25] - Connecting to assignment server
[16:03:26] - Successful: assigned to (0.0.0.0).
[16:03:26] + News From Folding@Home: Welcome to Folding@Home
[16:03:26] Work Unit has an invalid address.
[16:03:26] - Error: Attempt #1 to get work failed, and no other work to do.
Waiting before retry.

So that suggests I connected to the assignment server without a problem but that the server couldn't give me a work unit.
If there is some set of configurations I should be using I'd appreciate knowing what they are.

John Dean wrote:...If there is some set of configurations I should be using I'd appreciate knowing what they are.

You haven't posted what settings you are using, so how can we suggest what to change? Please post the first 30 lines of your fahlog.txt file.

Also, if you have manually edited the client.cfg file, you may have corrupted it to where there conflicting settings, and you won't ever get a new work unit.

For reference, using the defaults settings, and using no client switches will give you the widest chance of getting new work. The exception is with the SMP client where you need BigWUs turned on.

7im wrote:

John Dean wrote:...If there is some set of configurations I should be using I'd appreciate knowing what they are.

You haven't posted what settings you are using, so how can we suggest what to change? Please post the first 30 lines of your fahlog.txt file.

Also, if you have manually edited the client.cfg file, you may have corrupted it to where there conflicting settings, and you won't ever get a new work unit.

For reference, using the defaults settings, and using no client switches will give you the widest chance of getting new work. The exception is with the SMP client where you need BigWUs turned on.

fahlog.txt reads:

--- Opening Log file [January 24 12:57:45]


# Windows Graphical Edition ###################################################
###############################################################################

Folding@Home Client Version 5.03

http://folding.stanford.edu

###############################################################################
###############################################################################

Launch directory: C:\Program Files\Folding@Home


[12:57:45] - Ask before connecting: No
[12:57:45] - User name: dandy66 (Team 446)
[12:57:45] - User ID: 5F0C06901DDF8FEE
[12:57:45] - Machine ID: 1
[12:57:45]
[12:57:45] Loaded queue successfully.
[12:57:45] Initialization complete
[12:57:45] + Benchmarking ...
[12:57:48]
[12:57:48] + Processing work unit
[12:57:48] Core required: FahCore_79.exe
[12:57:49] Core found.
[12:57:49] Working on Unit 02 [January 24 12:57:49]
[12:57:49] + Working ...
[12:57:50]
[12:57:50] *------------------------------*
[12:57:50] Folding@Home Double Gromacs Core
[12:57:50] Version 1.91 (April 11, 2006)

I would have NO idea how to begin editing the client.cfg file
I hope I haven't given you the impression of someone who knows what he's doing in his computer. Nothing could be further from the truth.
I'm not too sure what 'client switches' are - I just tick the boxes that seem to need ticking or which people here suggest ought to be ticked. And if they say 'leave unticked' that's what I do

Per 7im's instructions, you have posted the first portion of FAHlog.txt. It looks just fine. I guess we're going to need the last couple of pages of FAHlog.txt to better understand the most recent activity.

Don't worry about command line switches or files that you have not editied. It's enough to know that you didn't do those things.

Something wasn't making sense to me, so I went back and read the whole thread again.

So if I understand the problem correctly, the fah client IS working for the most part. However, it will occasionally have problems picking up new work units. Sometimes you see the 0.0.0.0 error, and sometimes the client connects to a server, but can't download. And some minutes or hours later, the fah client will eventually get new work.

Assuming the above is an accurate description, then the fah client appears to be working as programmed.

The 0.0.0.0 error should be rare, as Stanford usually has a lot of CPU work units available. Recent work on the Stanford network may have contributed to this error lately. If the 0.0.0.0 error message happens a lot, then we will continue to investigate that particulare issue.

The connecting but not download is more common, but still not a big concern. Sometimes a particular server is busy, and how reliably the client can connect also depends on your network connection. (are you on dialup, DSL, Cable?) After several failed attempts to connect, the Assignment Server should signal the client to attempt a connection with a different server.

Less common, but possible, is that the Assignment server will not redirect the client to a different work server. Alternate servers may be down for maintenance or upgrades, out of work units, or even misprogrammed so no failover takes place. The client is then limited to one work server, and will eventually get new work when that server is less busy.

But the fah client and servers are programmed to handle all of these possible situations, and will continue attempting to get new work. There is little you can do to optimize this process. The mixed results tell me it is probably not your fah client having the issue. Bad settings, etc. typically cause a consistent failure. Since you have not added client switches, and you use the default client settings, and have a reliable net connection, then you are doing the best you can.