World Community Grid - View Thread - upload server: EOF on socket read

World Community Grid Forums

Category: Support

Forum: BOINC Agent Support

Thread: upload server: EOF on socket read

Quick Go »

No member browsing this thread

Thread Status: Active
Total posts in this thread: 5

[ ]

Author

This topic has been viewed 852 times and has 4 replies

Former Member
Cruncher
Joined: May 22, 2018
Post Count: 0
Status: Offline


upload server: EOF on socket read

My BOINC client has been trying to upload a particular result for over two days now. It's an error I've never seen before. Here are the messages:

Started upload of CMD2_1905-1XCD_A.clustersOccur-2V5Y_A.clustersOccur_22_1_0
[error] Error reported by file upload server: EOF on socket read : asked for 16382, got 0
Temporarily failed upload of CMD2_1905-1XCD_A.clustersOccur-2V5Y_A.clustersOccur_22_1_0: transient upload error

I've restart the client, rebooted my computer and reset my router and modem. Other results that have completed after this workunit, including ones for the same project (Help Cure Muscular Dystrophy - Phase 2), have uploaded fine.

Any ideas?

[Jun 15, 2011 5:44:48 PM]

Former Member
Cruncher
Joined: May 22, 2018
Post Count: 0
Status: Offline


Re: upload server: EOF on socket read

Hi,

On search find this error only ever being reported in the Active Research forums for CEP2 and FAAH. Maybe it's a question of the techs resetting a flag or deleting a file on the servers for a retry. Can you plz visit the stdoutdae.txt file or current BOINC Manager message/event log and see what happened the first time the upload was started.

ttyl.

[Jun 15, 2011 6:22:32 PM]

Former Member
Cruncher
Joined: May 22, 2018
Post Count: 0
Status: Offline


Re: upload server: EOF on socket read

On search find this error only ever being reported in the Active Research forums for CEP2 and FAAH. Maybe it's a question of the techs resetting a flag or deleting a file on the servers for a retry. Can you plz visit the stdoutdae.txt file or current BOINC Manager message/event log and see what happened the first time the upload was started.

Hi SekeRob,

Here's what I found in the log (FYI, I'm using Linux so it's in the /var/log/boinc.log) that was relevant to this workunit:

12-Jun-2011 23:57:14 [World Community Grid] Computation for task CMD2_1905-1XCD_A.clustersOccur-2V5Y_A.clustersOccur_22_1 finished
12-Jun-2011 23:57:16 [World Community Grid] Started upload of CMD2_1905-1XCD_A.clustersOccur-2V5Y_A.clustersOccur_22_1_0
12-Jun-2011 23:57:19 [---] Project communication failed: attempting access to reference site
12-Jun-2011 23:57:19 [World Community Grid] Temporarily failed upload of CMD2_1905-1XCD_A.clustersOccur-2V5Y_A.clustersOccur_22_1_0: can't resolve hostname
12-Jun-2011 23:57:19 [World Community Grid] Backing off 1 min 0 sec on upload of CMD2_1905-1XCD_A.clustersOccur-2V5Y_A.clustersOccur_22_1_0
13-Jun-2011 00:06:57 [World Community Grid] Started upload of CMD2_1905-1XCD_A.clustersOccur-2V5Y_A.clustersOccur_22_1_0
13-Jun-2011 00:07:01 [---] Received signal 15
13-Jun-2011 00:07:01 [---] Exit requested by user
13-Jun-2011 00:07:02 [---] [error] Deleting file CMD2_1905-1XCD_A.clustersOccur-2V5Y_A.clustersOccur_22_1_0 while in use
13-Jun-2011 00:07:12 [World Community Grid] Started upload of CMD2_1905-1XCD_A.clustersOccur-2V5Y_A.clustersOccur_22_1_0
13-Jun-2011 00:07:12 [World Community Grid] Restarting task E202394_430_C.26.C21H11N3OS.00554978.0.set1d06_0 using cep2 version 640
13-Jun-2011 00:07:12 [World Community Grid] Restarting task c4cw_target04_015119049_1 using c4cw version 641
13-Jun-2011 00:07:17 [World Community Grid] [error] Error reported by file upload server: EOF on socket read : asked for 16382, got 0
13-Jun-2011 00:07:17 [World Community Grid] Temporarily failed upload of CMD2_1905-1XCD_A.clustersOccur-2V5Y_A.clustersOccur_22_1_0: transient upload error
13-Jun-2011 00:07:17 [World Community Grid] Backing off 1 min 0 sec on upload of CMD2_1905-1XCD_A.clustersOccur-2V5Y_A.clustersOccur_22_1_0

A check of the projects directory does show a file CMD2_1905-1XCD_A.clustersOccur-2V5Y_A.clustersOccur_22_1_0 with 0 bytes.

Is it possible the result somehow got corrupted and my BOINC Client and the WCG server can't figure out what to do with it?

[Jun 16, 2011 2:52:47 AM]

Former Member
Cruncher
Joined: May 22, 2018
Post Count: 0
Status: Offline


Re: upload server: EOF on socket read

Hi,

Yes, was afraid of a corruption:

13-Jun-2011 00:07:01 [---] Received signal 15

I have the same thing sometimes when my wireless goes down on Linux and then the client trying to connect to the server [reason I'm setting my clients to a small daily connect window]. Little hope, think you best abort the result in the Tasks tab of BOINC Manager.

The log name is interesting... I've never looked in that path. What distro is this?

--//--

[Jun 16, 2011 3:00:08 AM]

Former Member
Cruncher
Joined: May 22, 2018
Post Count: 0
Status: Offline


Re: upload server: EOF on socket read

After seeing the messages in the log it makes sense. I'm on a wired connection but there could have been something else that caused the corruption in my particular case. No big deal. BOINC in general and WCG specifically are pretty stable nowadays so an occasional lost result is nothing to get in a twist about.

I'm currently running Fedora 14 with the BOINC Client from the Fedora repository. The log has been there since I switched to using the repo version starting with F12 or maybe F11, I can't remember.

[Jun 16, 2011 6:02:53 PM]

[ ]