| Index | Recent Threads | Unanswered Threads | Who's Active | Guidelines | Search |
 |
World Community Grid Forums
Category: Support
Forum: BOINC Agent Support
Thread: Uploads occasionally get stuck, "Retry now" ignored, BOINC 5.10.30 |
| No member browsing this thread |
|
Thread Status: Active Total posts in this thread: 7
|
|
| Author |
|
|
Rickjb
Veteran Cruncher Australia Joined: Sep 17, 2006 Post Count: 666 Status: Offline Project Badges: 
|
I'm running BOINC 5.10.30 (Win 32), Win 2kSP4, Internet via ADSL2+. Very occasionally, a file upload gets stuck. Clicking "Retry now" does not restart it. The most recent one was stuck at 0.0% for over 22 hours, ie the jam does not time out. At a whim, I tried "Activity >> Network activity suspended", then clicked "Retry now" on the Transfers tab. The jammed file uploaded immediately. Afterwards, I restored the "Network activity always available setting".
I think that BOINC's recovery from jammed uploads could be improved, ie so that this will occur automatically. The jams are probably caused by problems with my ADSL2+ connection, at one or more of the ISP, telephone exchange, Netgear router or Windows TCP/IP daemon levels. I run some bittorrent traffic at very restricted speeds, and this seems to produce the problems, which frequently affect web browsing in various erratic ways. I've been trying to fix that, but meanwhile I have exposed a weakness in BOINC. |
||
|
|
Sekerob
Ace Cruncher Joined: Jul 24, 2005 Post Count: 20043 Status: Offline |
Try 5.10.42 almost final release as that has a number of improvements on the comms front. WCG is preparing a new 'recommended' release which I believe is this version for windows. The work around to suspend network in BOINC completely and resume is good. Suspend and hitting update opens a line automatically for 5 minutes and refreshes a few bits.
----------------------------------------Async is returned to the fold and it has much better Vista shutdown and proxy handling e.g. in this release.
WCG
Please help to make the Forums an enjoyable experience for All! |
||
|
|
JmBoullier
Former Community Advisor Normandy - France Joined: Jan 26, 2007 Post Count: 3716 Status: Offline Project Badges: 
|
And in cases where all the above is not enough the next step is "simply close Boinc gently and restart it". Until now this has always been enough for me to solve such cases of idleness of one process or another.
----------------------------------------Cheers. Jean. |
||
|
|
Rickjb
Veteran Cruncher Australia Joined: Sep 17, 2006 Post Count: 666 Status: Offline Project Badges:
|
This bug is still present in BOINC 5.10.45 (32-bit Windows).
----------------------------------------Previously reported under Win 2k Pro SP4, this time XP Pro SP3. I recently had a single DDDT 400kb result file upload get stuck at 0kb transferred. Other BOINC network activity carried on normally while the upload was stuck. The task remained on the BOINC Tasks tab with Status "Uploading". "Retry now" on the Transfers tab, and "Suspend"/"Resume" on the Tasks tab were ignored. After I did "Network activity suspended" and a few seconds later "Network activity always available", from the Activity menu, the upload proceeded normally, and the result was validated. Relevant parts of the Messages log are appended below. Please communicate this problem to the BOINC programmers, and suggest that they concentrate on function rather than form. No "skins" and pretty progress-bars until they fix the bugs! WCG (presumably) has lots of crunchers who are not computer enthusiasts, and they are likely to drop out if things get stuck and they have to intervene. And Berkeley has a reputation to defend for bulletproof no-frills software (eg BSD Unix). Suggested approach to a fix: The problem is rare, not reproducible, and there may be more than 1 cause. Information that there is a problem is available: non-progressing transfer, other successful network activity since the stuck transfer started or made progress. Automatically take workaround actions as per manual procedure described above, rather than perhaps wasting lots of time hunting for the cause(s). ======== Affected WU and result: dddt0602h0469_ 100116_ 0-- rjb-q9450a Valid 08/03/2008 04:09:09 08/03/2008 18:36:42 2.56 62.9 / 72.5 Messages log: ## Lines starting with ## are comments that I have inserted 3/08/2008 10:15:53 PM|World Community Grid|Finished download of dddt0602h0474_100038_wcgrid.00049.dpf ## Task finishes: 3/08/2008 11:27:00 PM|World Community Grid|Computation for task dddt0602h0469_100116_0 finished 3/08/2008 11:27:00 PM|World Community Grid|Starting dddt0602h0470_100246_0 3/08/2008 11:27:00 PM|World Community Grid|Starting task dddt0602h0470_100246_0 using dddt version 606 3/08/2008 11:27:03 PM|World Community Grid|Started upload of dddt0602h0469_100116_0_0 ## Start of upload that gets stuck (file dddt0602h0469_100116_0_1): 3/08/2008 11:27:03 PM|World Community Grid|Started upload of dddt0602h0469_100116_0_1 3/08/2008 11:27:49 PM|World Community Grid|Finished upload of dddt0602h0469_100116_0_0 ## ... ## ... more network activity (deleted from this post) ## ... 4/08/2008 3:21:10 AM|World Community Grid|Finished upload of dddt0602h0471_100133_0_3 ## By here I had tried Transfers >> Retry now, and Tasks >> Suspend/Resume ## Then I did Activity >> "Network activity suspended", ## and a few seconds later, "Network activity always available": 4/08/2008 3:21:16 AM||Suspending network activity - user request 4/08/2008 3:21:26 AM||Resuming network activity 4/08/2008 3:21:26 AM|World Community Grid|Started upload of dddt0602h0469_100116_0_1 4/08/2008 3:21:43 AM|World Community Grid|Finished upload of dddt0602h0469_100116_0_1 ## Done 4/08/2008 4:01:56 AM|World Community Grid|Computation for task dddt0602h0472_100093_0 finished [Edit 2 times, last edit by Rickjb at Aug 4, 2008 12:53:19 PM] |
||
|
|
Sekerob
Ace Cruncher Joined: Jul 24, 2005 Post Count: 20043 Status: Offline |
We hear you and agree that bug fixes is the Numero Uno when scheduling programming, particular in the area of networking, graceful recovery AND the none effecting of computing itself as it does in instances (There's RPC redesign in the works for that).
----------------------------------------Until then, have you tried the "Do Network Communication" from the Advanced menu? It's little known but seems to kick things back in shape.
WCG
Please help to make the Forums an enjoyable experience for All! |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Until then, have you tried the "Do Network Communication" from the Advanced menu? It's little known but seems to kick things back in shape. Of course in BOINC 5.10.45 this is called "Retry Communications" and can be found under Advanced. Sekerob is using the new versin of the BOINC client which is not recommended for WCG at the moment. |
||
|
|
Sekerob
Ace Cruncher Joined: Jul 24, 2005 Post Count: 20043 Status: Offline |
 That's how little known it is.Yes, this is test driving 6.2.14, which has already been superseded by the next test iteration 6.2.15. Another sample of the consistency coming out of Berkeley.... without knowing why, the reasoning of this name change befuddles. Does not, as in the 5.10 versions, create a message log entry to confirm as positive feedback. Suspect if there was anything to communicate, it's considered all that's needed for a user. 
WCG
Please help to make the Forums an enjoyable experience for All! |
||
|
|
|