Don't use the image processing in the OCR plugin - only use the PDF processing. This issue seems to be related to a thread semaphore problem - I'll fix this, if I'll find the time. I plan to remove the image processing from this plugin until end of this year.
>Why does the documentation say >ASSP_OCRocrmaxprocesses should be less than the number of cpu cores? OCR (ImageMagick + tesseract) will use 100% of one core per concurrent processed image until it has finished. I expect a maximum of 0.05% better spam detection using the OCR-image processing - for an up to 95% higher CPU usage, this is toooo less. It is much more efficient to use the 'ASSP_AFCDetectSpamAttachRe' in the ASSP_AFC plugin. Thomas Von: "Dirk Kulmsee" <d.kulm...@netgroup.de> An: <assp-test@lists.sourceforge.net> Datum: 02.09.2014 16:27 Betreff: [Assp-test] ASSP_OCR & stuck workers question Hi, I am currently running ASSP 2.4.4 (14241) on Debian Linux with Perl 5.20. The ASSP_OCR module is 2.18. I had all worker processes stuck in ASSP_OCR one by one: 2014-09-02 10:59:21 [Main_Thread] Info: Loop in Worker_1 was not active for 461 seconds 2014-09-02 10:59:21 [Main_Thread] Info: Worker_1 : last sigoff in ASSP_OCR, /opt/assp/Plugins/ASSP_OCR.pm, 282, main::sigoffTry, 1, , , at 14-2-8 10:51:40 1409647900.23592 - 282 2014-09-02 10:59:21 [Main_Thread] Info: Worker_1 : last sigon in main, sub main::URIBLok, 15, main::URIBLok_Run, 1, , , at 14-2-8 10:51:40 1409647900.2248 - 272 2014-09-02 10:59:21 [Main_Thread] Info: Worker_1 : last action was : call Plugin ASSP_OCR with 2014-09-02 10:59:21 [Main_Thread] Warning: try to terminate inactive/stucking Worker_1 2014-09-02 11:19:26 [Main_Thread] Info: Loop in Worker_2 was not active for 466 seconds 2014-09-02 11:19:26 [Main_Thread] Info: Worker_2 : last sigoff in ASSP_OCR, /opt/assp/Plugins/ASSP_OCR.pm, 282, main::sigoffTry, 1, , , at 14-2-8 11:11:40 1409649100.27879 - 282 2014-09-02 11:19:26 [Main_Thread] Info: Worker_2 : last sigon in main, sub main::URIBLok, 15, main::URIBLok_Run, 1, , , at 14-2-8 11:11:40 1409649100.26713 - 241 2014-09-02 11:19:26 [Main_Thread] Info: Worker_2 : last action was : call Plugin ASSP_OCR with 2014-09-02 11:19:26 [Main_Thread] Warning: try to terminate inactive/stucking Worker_2 2014-09-02 11:36:11 [Main_Thread] Info: Loop in Worker_3 was not active for 271 seconds 2014-09-02 11:36:11 [Main_Thread] Info: Worker_3 : last sigoff in ASSP_OCR, /opt/assp/Plugins/ASSP_OCR.pm, 282, main::sigoffTry, 1, , , at 14-2-8 11:31:40 1409650300.57724 - 282 2014-09-02 11:36:11 [Main_Thread] Info: Worker_3 : last sigon in main, sub main::URIBLok, 15, main::URIBLok_Run, 1, , , at 14-2-8 11:31:40 1409650300.56076 - 241 2014-09-02 11:36:11 [Main_Thread] Info: Worker_3 : last action was : call Plugin ASSP_OCR with 2014-09-02 11:36:11 [Main_Thread] Warning: try to terminate inactive/stucking Worker_3 2014-09-02 13:49:57 [Main_Thread] Info: Loop in Worker_4 was not active for 196 seconds 2014-09-02 13:49:57 [Main_Thread] Info: Worker_4 : last sigoff in ASSP_OCR, /opt/assp/Plugins/ASSP_OCR.pm, 282, main::sigoffTry, 1, , , at 14-2-8 13:46:41 1409658401.38248 - 282 2014-09-02 13:49:57 [Main_Thread] Info: Worker_4 : last sigon in main, sub main::URIBLok, 15, main::URIBLok_Run, 1, , , at 14-2-8 13:46:41 1409658401.36525 - 241 2014-09-02 13:49:57 [Main_Thread] Info: Worker_4 : last action was : call Plugin ASSP_OCR with 2014-09-02 13:49:57 [Main_Thread] Warning: try to terminate inactive/stucking Worker_4 Later I found a live example for this. A simple email status report containing four little PNG icons stuck the worker process, leaving log lines like these: 2014-09-02 13:59:26 m1-59166-11063 [Worker_1] [Plugin] 88.198.3.4 [OIP: 81.209.171.97] <server2...@someone.de> to: al...@mydomain.de ASSP_OCR: (att) file text1.ecelp9600 found in mime part 1 2014-09-02 13:59:26 m1-59166-11063 [Worker_1] [Plugin] 88.198.3.4 [OIP: 81.209.171.97] <server2...@someone.de> to: al...@mydomain.de ASSP_OCR: (att) file logo.png found in mime part 2 2014-09-02 13:59:26 m1-59166-11063 [Worker_1] [Plugin] 88.198.3.4 [OIP: 81.209.171.97] <server2...@someone.de> to: al...@mydomain.de ASSP_OCR: processing (attatched) file logo.png 2014-09-02 13:59:26 m1-59166-11063 [Worker_1] [Plugin] 88.198.3.4 [OIP: 81.209.171.97] <server2...@someone.de> to: al...@mydomain.de ASSP_OCR: (att) file warning.png found in mime part 3 2014-09-02 13:59:26 m1-59166-11063 [Worker_1] [Plugin] 88.198.3.4 [OIP: 81.209.171.97] <server2...@someone.de> to: al...@mydomain.de ASSP_OCR: processing (attatched) file warning.png 2014-09-02 13:59:26 m1-59166-11063 [Worker_1] [Plugin] 88.198.3.4 [OIP: 81.209.171.97] <server2...@someone.de> to: al...@mydomain.de ASSP_OCR: (att) file success.png found in mime part 4 2014-09-02 13:59:26 m1-59166-11063 [Worker_1] [Plugin] 88.198.3.4 [OIP: 81.209.171.97] <server2...@someone.de> to: al...@mydomain.de ASSP_OCR: processing (attatched) file success.png 2014-09-02 13:59:26 m1-59166-11063 [Worker_1] [Plugin] 88.198.3.4 [OIP: 81.209.171.97] <server2...@someone.de> to: al...@mydomain.de ASSP_OCR: (att) file error.png found in mime part 5 2014-09-02 13:59:26 m1-59166-11063 [Worker_1] [Plugin] 88.198.3.4 [OIP: 81.209.171.97] <server2...@someone.de> to: al...@mydomain.de ASSP_OCR: processing (attatched) file error.png I looked into the config for ASSP_OCR and found ASSP_OCRocrmaxprocesses set to its default value of three. Here comes the funny part: When ASSP_OCRocrmaxprocesses is set to 3, the worker gets stuck as soon as it hits png #4 When ASSP_OCRocrmaxprocesses is set to 1, the worker gets stuck as soon as it hits png #2 When ASSP_OCRocrmaxprocesses is set to 10, this email gets through and I have no stuck worker processes since (at least for the last two hours :) ). Can anyone confirm this? Could it be, that ASSP_OCR goes mad when it finds "ASSP_OCRocrmaxprocesses +1 " images? Why does the documentation say ASSP_OCRocrmaxprocesses should be less than the number of cpu cores? Best regards Dirk ------------------------------------------------------------------------------ Slashdot TV. Video for Nerds. Stuff that matters. http://tv.slashdot.org/ _______________________________________________ Assp-test mailing list Assp-test@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/assp-test DISCLAIMER: ******************************************************* This email and any files transmitted with it may be confidential, legally privileged and protected in law and are intended solely for the use of the individual to whom it is addressed. This email was multiple times scanned for viruses. There should be no known virus in this email! ******************************************************* ------------------------------------------------------------------------------ Slashdot TV. Video for Nerds. Stuff that matters. http://tv.slashdot.org/ _______________________________________________ Assp-test mailing list Assp-test@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/assp-test
