| Index | Recent Threads | Unanswered Threads | Who's Active | Guidelines | Search |
 |
World Community Grid Forums
Category: Active Research
Forum: Africa Rainfall Project
Thread: Error - finish file present too long |
| No member browsing this thread |
|
Thread Status: Active Total posts in this thread: 7
|
|
| Author |
|
|
gb009761
Master Cruncher Scotland Joined: Apr 6, 2005 Post Count: 2982 Status: Offline Project Badges: 
|
Hi WCG Techies,
----------------------------------------A WU you may want to have a 'nosy' at (as I haven't personally seen an error of this type before - nor did I spot a thread on the ARP1 forum of the same ilk)... ARP1_0014879_010_1-- FrancisT550 Error 4/30/20 17:03:38 5/3/20 21:17:18 32.09 / 38.53 687.6 / 0.0 Result Log Result Name: ARP1_ 0014879_ 010_ 1-- <core_client_version>7.14.2</core_client_version> <![CDATA[ <stderr_txt> INFO: Initializing INFO: No state to restore. Start from the beginning. Starting WRFMain [10:41:20] INFO: Checkpoint taken at 2018-07-21_06:00:00 [16:48:06] INFO: Checkpoint taken at 2018-07-21_12:00:00 [22:33:10] INFO: Checkpoint taken at 2018-07-21_18:00:00 [03:18:48] INFO: Checkpoint taken at 2018-07-22_00:00:00 [06:56:46] INFO: Checkpoint taken at 2018-07-22_06:00:00 [12:23:09] INFO: Checkpoint taken at 2018-07-22_12:00:00 [17:58:54] INFO: Checkpoint taken at 2018-07-22_18:00:00 [21:44:33] INFO: Checkpoint taken at 2018-07-23_00:00:00 INFO: Simulation complete compressing output. 21:55:43 (12612): called boinc_finish(0) </stderr_txt> <message> finish file present too long</message> ]]>  [Edit 1 times, last edit by gb009761 at May 4, 2020 4:29:15 AM] |
||
|
|
Jim1348
Veteran Cruncher USA Joined: Jul 13, 2009 Post Count: 1066 Status: Offline Project Badges: 
|
</stderr_txt> <message> finish file present too long</message> ]]> I have seen that mentioned on other forums. It usually means that BOINC can't access the disk fast enough to send back the results. Try upgrading to the latest version of BOINC, or reducing the number of cores you are running. I use a fast SSD, with a large (2 GB or more) write-cache, and don't recall seeing it myself. That is on Linux, which has a built-in write cache; you just need to adjust the size. You can buy PrimoCache for Windows, or use the utility that comes with Samsung or Crucial SSDs, which includes caching software. |
||
|
|
Mike.Gibson
Ace Cruncher England Joined: Aug 23, 2007 Post Count: 12360 Status: Offline Project Badges: 
|
ARP files are massive in comparison to other projects. Did you upload more than one at the same time? How fast is your broadband connection?
It is not recommended to run arp on mare than half the threads that you have. Smaller units like mcm can be run on the other half. More than half and they slow down. Even on half, it is best to prevent 2 arp checkpointing at the same time, let alone uploading. Mike |
||
|
|
Jim1348
Veteran Cruncher USA Joined: Jul 13, 2009 Post Count: 1066 Status: Offline Project Badges:
|
While waiting for three CPDNs to finish on my Ryzen 3900X (the HadAM4 are also rather heavy on the writes) I wondered if I could get some ARPs. I ended up with nine ARP, and the other cores on SCC and MIP (24 cores in all). It was not a problem with my SSD and cache on Ubuntu 18.04.4.
|
||
|
|
nanoprobe
Master Cruncher Classified Joined: Aug 29, 2008 Post Count: 2998 Status: Offline Project Badges: 
|
Hi WCG Techies, A WU you may want to have a 'nosy' at (as I haven't personally seen an error of this type before - nor did I spot a thread on the ARP1 forum of the same ilk)... ARP1_0014879_010_1-- FrancisT550 Error 4/30/20 17:03:38 5/3/20 21:17:18 32.09 / 38.53 687.6 / 0.0 Result Log Result Name: ARP1_ 0014879_ 010_ 1-- <core_client_version>7.14.2</core_client_version> <![CDATA[ <stderr_txt> INFO: Initializing INFO: No state to restore. Start from the beginning. Starting WRFMain [10:41:20] INFO: Checkpoint taken at 2018-07-21_06:00:00 [16:48:06] INFO: Checkpoint taken at 2018-07-21_12:00:00 [22:33:10] INFO: Checkpoint taken at 2018-07-21_18:00:00 [03:18:48] INFO: Checkpoint taken at 2018-07-22_00:00:00 [06:56:46] INFO: Checkpoint taken at 2018-07-22_06:00:00 [12:23:09] INFO: Checkpoint taken at 2018-07-22_12:00:00 [17:58:54] INFO: Checkpoint taken at 2018-07-22_18:00:00 [21:44:33] INFO: Checkpoint taken at 2018-07-23_00:00:00 INFO: Simulation complete compressing output. 21:55:43 (12612): called boinc_finish(0) </stderr_txt> <message> finish file present too long</message> ]]> The "finish file present too long" problem is caused by the BOINC client. See this bug report: https://github.com/BOINC/boinc/issues/3017 and references therein.
In 1969 I took an oath to defend and protect the U S Constitution against all enemies, both foreign and Domestic. There was no expiration date.
  |
||
|
|
gb009761
Master Cruncher Scotland Joined: Apr 6, 2005 Post Count: 2982 Status: Offline Project Badges:
|
Thanks folks for replying. As alluded to above and in that thread nanoprobe linked to, I think "swap thrashing" is the most likely cause. Very unfortunate on these mega WU's.
----------------------------------------Just to clarify, on both my machines, I've got a quad core, and generally only allow 2 ARP1 WU's to run at the same time (very occasionally - three, if they're running the risk of missing their deadline, when I ensure that there'll be no contention with checkpointing). The machine this happened on, is running Win 8.1 - which seems to have issues with the swap file (I don't know why - I've looked, but can't see anything obvious), whilst my other machine (Win7), hasn't had any kind of issue at all (other than being slower than my newer machine). As this is the first time it's happened to me (and, as there's no other thread in the ARP1 forum), hopefully it's a rare event - I've certainly been able to get through over 160 other ARP1 WU's with no issues such as this. |
||
|
|
Mike.Gibson
Ace Cruncher England Joined: Aug 23, 2007 Post Count: 12360 Status: Offline Project Badges:
|
There is just a possibility that your broadband speed was reduced on this occasion. The server might have got impatient and shut you off. They are big uploads.
Mike |
||
|
|
|