Cannot connect to get jobs

[aardWolf]

I installed it on an Ubuntu VM and am getting Connection Refused from 171.64.65.121:80. I was able to get it installed in a Windows 7 VM, and am getting the following error:

Code: Select all

13:59:06:Saving configuration to config.xml
13:59:06:<config>
13:59:06:  <!-- Network -->
13:59:06:  <proxy v=':8080'/>
13:59:06:
13:59:06:  <!-- User Information -->
13:59:06:  <passkey v='********************************'/>
13:59:06:  <user v='LawsonCulver'/>
13:59:06:
13:59:06:  <!-- Folding Slots -->
13:59:06:  <slot id='0' type='CPU'/>
13:59:06:</config>
13:59:14:WARNING:WU00:FS00:Failed to get ID from '171.67.108.200:8080': Failed to connect to 171.67.108.200:8080: A connection attempt failed because the connected party did not properly respond after a period of time, or established connection failed because connected host has failed to respond.
13:59:14:WU00:FS00:Connecting to 171.64.65.121:80
13:59:16:WARNING:WU00:FS00:Failed to get ID from '171.64.65.121:80': Failed to connect to 171.64.65.121:80: No connection could be made because the target machine actively refused it.
13:59:16:ERROR:WU00:FS00:Exception: Could not get an assignment ID
13:59:16:WU00:FS00:Connecting to 171.67.108.200:8080
13:59:37:WARNING:WU00:FS00:Failed to get ID from '171.67.108.200:8080': Failed to connect to 171.67.108.200:8080: A connection attempt failed because the connected party did not properly respond after a period of time, or established connection failed because connected host has failed to respond.
13:59:37:WU00:FS00:Connecting to 171.64.65.121:80
13:59:38:WARNING:WU00:FS00:Failed to get ID from '171.64.65.121:80': Failed to connect to 171.64.65.121:80: No connection could be made because the target machine actively refused it.
13:59:38:ERROR:WU00:FS00:Exception: Could not get an assignment ID
13:59:54:Saving configuration to config.xml
13:59:54:<config>
13:59:54:  <!-- Network -->
13:59:54:  <proxy v=':8080'/>
13:59:54:
13:59:54:  <!-- User Information -->
13:59:54:  <passkey v='********************************'/>
13:59:54:  <user v='LawsonCulver'/>
13:59:54:
13:59:54:  <!-- Folding Slots -->
13:59:54:  <slot id='0' type='CPU'/>
13:59:54:</config>

I installed Chrome to try the NaCL client, and I got: "Warning: Work Unit assignment failed". I don't see any kind of logs that might give me more insight into that one...

[7im]

Just out of interests: could you install
A) a virtual Linux under VM Fusion and see if you get that one folding, to exclude local Mac OSX issues
B) install the NaCl client as they use complete different server (to my understanding);
Of course same donor ID, ..., and passkey for both options

Not a Mac issue, proven by tether working, so no need to try Linux.
Not a server issue, proven by tether working, but do try the NaCl client to help diagnose your networking issue. If the chrome browser gets through, but fah, which uses the same ports as a browser does not get through, then it would confirm some AV, FW, or network/local proxy issue.

[aardWolf]

Just out of interests: could you install
A) a virtual Linux under VM Fusion and see if you get that one folding, to exclude local Mac OSX issues
B) install the NaCl client as they use complete different server (to my understanding);
Of course same donor ID, ..., and passkey for both options

Not a Mac issue, proven by tether working, so no need to try Linux.
Not a server issue, proven by tether working, but do try the NaCl client to help diagnose your networking issue. If the chrome browser gets through, but fah, which uses the same ports as a browser does not get through, then it would confirm some AV, FW, or network/local proxy issue.

The NaCl client didn't work either...

[bollix47]

Try this:

Open Chrome and start up the NaCl client.
Press F12 and click on the Console tab.

The log may add some pieces to the puzzle.

[aardWolf]

Try this:

Open Chrome and start up the NaCl client.
Press F12 and click on the Console tab.

The log may add some pieces to the puzzle.

The NaCL client is only trying the same server over and over on port 8080. It hasn't attempted to connect to anything on port 80 since I've been watching it. This will never work on my network.

Code: Select all

DEBUG: Status: downloading: Downloading a work unit. main.js:73
DEBUG: Status: downloading: Downloading a work unit. main.js:73
Failed to load resource: net::ERR_CONNECTION_TIMED_OUT http://143.89.28.86:8080/api/assign?0.6672185291536152
DEBUG: Warning: Work Unit assignment failed, retrying main.js:73
DEBUG: Status: waiting: Waiting to retry work server assignment main.js:73
DEBUG: Status: downloading: Requesting a work server assignment. main.js:73
DEBUG: Status: downloading: Requesting a work server assignment. main.js:73
DEBUG: WS: {"client_id":"0xf434e05e533d6ca2","threads":24,"version":"8.1.0","type":"NACL","os":"NACL","user":"LawsonCulver","team":"3446","passkey":"*CENSORED*","ts":"2014-04-03T15:01:16Z","ws":"143.89.28.86","project":2981} main.js:73
DEBUG: Status: downloading: Downloading a work unit. main.js:73
DEBUG: Status: downloading: Downloading a work unit. main.js:73
OPTIONS http://143.89.28.86:8080/api/assign?0.5876986500807106 net::ERR_CONNECTION_TIMED_OUT jquery-1.10.2.min.js:4
DEBUG: Warning: Work Unit assignment failed, retrying main.js:73
DEBUG: Status: waiting: Waiting to retry work server assignment 

[7im]

ERR_CONNECTION_TIMED_OUT

This concerns me. The NaCl work units are extra small. They would even work on dial-up if the client still supported that.

What kind of network connection/ISP are you using again? What bandwidth? Heavily loaded? Experiencing a lot of packet loss? Dropped Pings?

[aardWolf]

ERR_CONNECTION_TIMED_OUT

This concerns me. The NaCl work units are extra small. They would even work on dial-up if the client still supported that.

What kind of network connection/ISP are you using again? What bandwidth? Heavily loaded? Experiencing a lot of packet loss? Dropped Pings?

Don't be concerned about that error message. The security guy here has blocked port 8080 outgoing so that it just times out. My client in the past has only been able to get work from a server on port 80. Because the NaCl client never attempts port 80, I won't know if it would have worked.

Everything worked for months with our current network setup... It's only the past few weeks that it stopped working even over port 80.

[bruce]

So much of your original problem seems to depend on what obstacles/protections are being placed by your employer that I'm not sure how much help we can offer beyond what has already been suggested. The fact that your phone tether works confirms that the problem is not server-based.

As far as it possibly being client-based, consider each of the following: The V6 client always tried port 8080 first and if it failed, then tried port 80. V7 started out the same way but I think it has changed recently and it probably will always try both if necessary. I'm not sure what NaCl does, but if we confirm that it uses only 8080, we can request an enhancement to use port 80 as a backup.

Does (almost) every connection to port 80 work as expected or is that still unreliable?

[aardWolf]

So much of your original problem seems to depend on what obstacles/protections are being placed by your employer that I'm not sure how much help we can offer beyond what has already been suggested. The fact that your phone tether works confirms that the problem is not server-based.

As far as it possibly being client-based, consider each of the following: The V6 client always tried port 8080 first and if it failed, then tried port 80. V7 started out the same way but I think it has changed recently and it probably will always try both if necessary. I'm not sure what NaCl does, but if we confirm that it uses only 8080, we can request an enhancement to use port 80 as a backup.

Does (almost) every connection to port 80 work as expected or is that still unreliable?

NaCl only uses port 8080 (or that's all it attempted for several hours yesterday).

Every connection to port 8080 fails with a timeout. The connection attempts to port 80 tend to have two possible results based on the particular server. Either:
1. No works units available
2. Access denied (503 error)

Interestingly enough, my client has no problems uploading the results.

Tethering (171.64.65.124)
:8080 works
:80 is never attempted (because :8080 works)
upload works

Ethernet (171.64.65.124)
:8080 times out
:80 gives no work units
upload works

I can't help but think that using a proxy server would help, if I could find one that is usable.

I think I could use SSH tunneling if I revert to the older client, because it uses a hostname instead of an IP to get work.

[bruce]

I've confirmed that NaCl, in its present form, does not use port 80 so we can drop it from the list of possible solutions for the time being.

A proxy server to bypass the proxy server that's blocking your assess? Maybe, but I'm not sure how.

SSH tunneling? again I'm not sure how. The FAH Client uses only one of two hostnames, both starting with "assign..." One Assignment Server is associated with port 8080 and the other is associated with port 80. Since for you, port 8080 is blocked, the connection to the first AS will always fail and the client will always use the secondary AS.

It's the job of the AS to choose an explicit IP address for one of the Work Servers. At that point, hostnames are never used. An explicit address is used for both for downloading and uploading.

The server at 171.64.65.124 will always be referenced either by 171.64.65.124:8080 (if you were directed to it by the first AS) or by 171.64.65.124:80 (if you were directed to it by the second AS. Nothing can be downloaded until contact with WS is established using the explicit address that was assigned by the AS.

Incidentally, that particular server has a limited number of WUs. At any particular time, it may or may not have WUs for you to download, though you will still be able to return completed WUs to it if that's where the originated. If it happens to be out of WUs, the AS will assign you to some other server that does have WUs for you. In the unlikely event that no WUs fit your criteria, you'll be given an Empty Work Server Assignment error. You never get to obtain a new WU using a hostname nor can you choose which Work Server will deliver one.

[aardWolf]

I was thinking possibly some combination of ipfw and tunneling, but that won't work. As you stated, the initial server assigns me a work server with an explicit IP. I guess for now I'll just tether my phone a few times a day. I sure would like to be able to set it and forget it like the one at home...

[bruce]

I'd like to be able to figure out why FAHClient is not working correctly through port 80 but so far, I can't explain it.

I predict that a future version of the NaCl client will have the same logic that has worked for others in a similar situation so understanding your situation is a challenge. Unfortunately during early client testing, a configuration like yours gets (almost?) no testing since it's rather rare.

[aardWolf]

I have quite an interesting revelation. I let my computer at work attempt to connect to 171.67.108.200:8080 (which I knew would time out.) I immediately switched my Internet connection over to tethering from my phone. When 8080 timed out, it connected to 171.64.65.121:80 and got "Empty work server assignment". It then connected to 8080 and got an assignment.

So, that's pretty compelling evidence to me that the port 80 servers don't have any work units for my machine (or just aren't working), but port 8080 does. If I could get work units available on port 80, then I could skip this whole "tether my iPhone" step.

Code: Select all

21:01:37:WARNING:WU00:FS00:Failed to get assignment from '171.67.108.200:8080': Failed to connect to 171.67.108.200:8080: Operation timed out
21:01:37:WU00:FS00:Connecting to 171.64.65.121:80
21:01:37:WARNING:WU00:FS00:Failed to get assignment from '171.64.65.121:80': Empty work server assignment
21:01:37:ERROR:WU00:FS00:Exception: Could not get an assignment
21:01:38:WU00:FS00:Connecting to 171.67.108.200:8080
21:01:39:WU00:FS00:Assigned to work server 171.64.65.124
21:01:39:WU00:FS00:Requesting new work unit for slot 00: READY cpu:24 from 171.64.65.124
21:01:39:WU00:FS00:Connecting to 171.64.65.124:8080
21:01:40:WU00:FS00:Downloading 808.61KiB
21:01:42:WU00:FS00:Download complete
21:01:42:WU00:FS00:Received Unit: id:00 state:DOWNLOAD error:NO_ERROR project:9007 run:496 clone:1 gen:30 core:0xa4 unit:0x00000020664f2de4533b357f8f273cb1
21:01:42:WU00:FS00:Starting
21:01:42:WU00:FS00:Running FahCore: /usr/bin/FAHCoreWrapper "/Library/Application Support/FAHClient/cores/www.stanford.edu/~pande/OSX/AMD64/Core_a4.fah/FahCore_a4" -dir 00 -suffix 01 -version 704 -lifeline 37608 -checkpoint 15 -np 24
21:01:42:WU00:FS00:Started FahCore on PID 71279
21:01:42:WU00:FS00:Core PID:71280
21:01:42:WU00:FS00:FahCore 0xa4 started

[7im]

Add support for standard port 80 - Ticket#57

[aardWolf]

This is the regular client that isn't working on port 80.