Thunderbird locks. [Was SUSE-SU-2022:3800-1: important: Security update for MozillaThunderbird]
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 On Thursday, 2022-10-27 at 18:22 +0200, opensuse-security@opensuse.org wrote: Thunderbird goes to 120% CPU when reading some Usenet groups. It can not even repaint the window. Previous version did the same, but worse. Only solution is to kill it externally and restart it crossing fingers. LOG from local proxy: <7.7> 2022-10-27T21:44:55.490871+02:00 Telcontar leafnode 574 - - markinterest: sci.electronics.design touched ctime <7.7> 2022-10-27T21:44:55.490954+02:00 Telcontar leafnode 574 - - >221 user-agent: header (from article files) for postings 65631-67059: <7.7> 2022-10-27T21:45:08.509136+02:00 Telcontar leafnode 574 - - <ARTICLE 67034 <7.7> 2022-10-27T21:45:08.509565+02:00 Telcontar leafnode 574 - - markinterest: sci.electronics.design touched ctime <7.7> 2022-10-27T21:45:08.509648+02:00 Telcontar leafnode 574 - - >220 67034 <61b3fa93-8f7f-4231-afa8-7a67c9f31086n@googlegroups.com> article retrieved - text follows <7.7> 2022-10-27T21:45:11.317053+02:00 Telcontar leafnode 574 - - <GROUP rec.arts.tv <7.7> 2022-10-27T21:45:11.317154+02:00 Telcontar leafnode 574 - - marked group rec.arts.tv interesting <7.7> 2022-10-27T21:45:11.317224+02:00 Telcontar leafnode 574 - - markinterest: rec.arts.tv touched ctime <7.7> 2022-10-27T21:45:11.317278+02:00 Telcontar leafnode 574 - - >211 219323 3 219325 rec.arts.tv group selected <7.7> 2022-10-27T21:45:11.333349+02:00 Telcontar leafnode 574 - - <XOVER 3-219325 <7.7> 2022-10-27T21:45:11.333476+02:00 Telcontar leafnode 574 - - markinterest: rec.arts.tv touched ctime <7.7> 2022-10-27T21:45:11.562841+02:00 Telcontar leafnode 574 - - >224 Overview information for postings 3-219325: <7.7> 2022-10-27T21:45:51.316135+02:00 Telcontar leafnode 574 - - <XHDR newsgroups 3-219325 <7.7> 2022-10-27T21:45:51.316280+02:00 Telcontar leafnode 574 - - markinterest: rec.arts.tv touched ctime <7.7> 2022-10-27T21:45:51.316329+02:00 Telcontar leafnode 574 - - >221 newsgroups: header (from article files) for postings 3-219325: <7.7> 2022-10-27T21:45:53.729587+02:00 Telcontar leafnode 574 - - <XHDR user-agent 3-219325 <7.7> 2022-10-27T21:45:53.729703+02:00 Telcontar leafnode 574 - - markinterest: rec.arts.tv touched ctime <7.7> 2022-10-27T21:45:53.729746+02:00 Telcontar leafnode 574 - - >221 user-agent: header (from article files) for postings 3-219325: - -- Cheers, Carlos E. R. (from openSUSE 15.3 x86_64 at Telcontar) -----BEGIN PGP SIGNATURE----- iHoEARECADoWIQQZEb51mJKK1KpcU/W1MxgcbY1H1QUCY1rhfxwccm9iaW4ubGlz dGFzQHRlbGVmb25pY2EubmV0AAoJELUzGBxtjUfVeWkAoIXjyRiSs1pHOXMSJKBq udBMre+VAJ96FucaPsaDo/ctyVKhZ4AMEJqy/Q== =zuSX -----END PGP SIGNATURE-----
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 On Thursday, 2022-10-27 at 21:52 +0200, Carlos E. R. wrote:
On Thursday, 2022-10-27 at 18:22 +0200, opensuse-security@opensuse.org wrote:
Thunderbird goes to 120% CPU when reading some Usenet groups. It can not even repaint the window. Previous version did the same, but worse. Only solution is to kill it externally and restart it crossing fingers.
LOG from local proxy:
<7.7> 2022-10-27T21:45:11.562841+02:00 Telcontar leafnode 574 - - >224 Overview information for postings 3-219325: <7.7> 2022-10-27T21:45:51.316135+02:00 Telcontar leafnode 574 - - <XHDR newsgroups 3-219325 <7.7> 2022-10-27T21:45:51.316280+02:00 Telcontar leafnode 574 - - markinterest: rec.arts.tv touched ctime <7.7> 2022-10-27T21:45:51.316329+02:00 Telcontar leafnode 574 - - >221 newsgroups: header (from article files) for postings 3-219325: <7.7> 2022-10-27T21:45:53.729587+02:00 Telcontar leafnode 574 - - <XHDR user-agent 3-219325 <7.7> 2022-10-27T21:45:53.729703+02:00 Telcontar leafnode 574 - - markinterest: rec.arts.tv touched ctime <7.7> 2022-10-27T21:45:53.729746+02:00 Telcontar leafnode 574 - - >221 user-agent: header (from article files) for postings 3-219325:
Log continues on kill: <7.7> 2022-10-27T21:52:39.557574+02:00 Telcontar leafnode 574 - - Client timeout, disconnecting. <7.3> 2022-10-27T21:52:39.558744+02:00 Telcontar leafnode 31383 - - Cannot write to client: Broken pipe <7.7> 2022-10-27T21:52:39.558801+02:00 Telcontar leafnode 31383 - - Client timeout, disconnecting. - -- Cheers, Carlos E. R. (from openSUSE 15.3 x86_64 at Telcontar) -----BEGIN PGP SIGNATURE----- iHoEARECADoWIQQZEb51mJKK1KpcU/W1MxgcbY1H1QUCY1riKxwccm9iaW4ubGlz dGFzQHRlbGVmb25pY2EubmV0AAoJELUzGBxtjUfVmdkAoIVy+cAHs8SdlXvvsW7o btpdcC+vAJ9UuFTJTgSmJwoeemflZW5QHC3aow== =u2mT -----END PGP SIGNATURE-----
On 10/27/22 14:55, Carlos E. R. wrote:
Log continues on kill:
<7.7> 2022-10-27T21:52:39.557574+02:00 Telcontar leafnode 574 - - Client timeout, disconnecting. <7.3> 2022-10-27T21:52:39.558744+02:00 Telcontar leafnode 31383 - - Cannot write to client: Broken pipe <7.7> 2022-10-27T21:52:39.558801+02:00 Telcontar leafnode 31383 - - Client timeout, disconnecting.
I'm not sure of the cause, but I've seen similar issues when the tbird index get screwed up. Generally a "compact" will cause the indexes to be regenerated. You can also simply delete the index files in the usegroup folder and it will be regenerated. Worth a shot. A search of thunderbird rebuild index will turn up the specific index file names -- which escapes me at the moment. -- David C. Rankin, J.D.,P.E.
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 On Thursday, 2022-10-27 at 15:30 -0500, David C. Rankin wrote:
On 10/27/22 14:55, Carlos E. R. wrote:
Log continues on kill:
<7.7> 2022-10-27T21:52:39.557574+02:00 Telcontar leafnode 574 - - Client timeout, disconnecting. <7.3> 2022-10-27T21:52:39.558744+02:00 Telcontar leafnode 31383 - - Cannot write to client: Broken pipe <7.7> 2022-10-27T21:52:39.558801+02:00 Telcontar leafnode 31383 - - Client timeout, disconnecting.
I'm not sure of the cause, but I've seen similar issues when the tbird index get screwed up. Generally a "compact" will cause the indexes to be regenerated. You can also simply delete the index files in the usegroup folder and it will be regenerated. Worth a shot. A search of thunderbird rebuild index will turn up the specific index file names -- which escapes me at the moment.
Repair folder sends TH into 100% CPU load and has to be killed again. As can be seen in the log, TH fails to read from the server, it locks or crashes. <7.7> 2022-10-27T22:37:16.923812+02:00 Telcontar leafnode 7470 - - config: debugmode is 1 <7.7> 2022-10-27T22:37:16.923944+02:00 Telcontar leafnode 7470 - - config: maxfetch is 5000 <7.7> 2022-10-27T22:37:16.924009+02:00 Telcontar leafnode 7470 - - config: maxage is 0 <7.7> 2022-10-27T22:37:16.924082+02:00 Telcontar leafnode 7470 - - config: postings have max. 500000 bytes <7.7> 2022-10-27T22:37:16.924140+02:00 Telcontar leafnode 7470 - - config: timeout_long is 100 days <7.7> 2022-10-27T22:37:16.924220+02:00 Telcontar leafnode 7470 - - config: timeout_fetchnews is 90 seconds <7.7> 2022-10-27T22:37:16.924278+02:00 Telcontar leafnode 7470 - - config: newsadmin is postmaster@telcontar.valinor <7.7> 2022-10-27T22:37:16.924368+02:00 Telcontar leafnode 7470 - - config: allow_8bit_headers is 1 <7.6> 2022-10-27T22:37:16.924444+02:00 Telcontar leafnode 7470 - - connect from localhost (::ffff:127.0.0.1) to localhost (::ffff:127.0.0.1) (my fqdn: Telcontar.valinor) <7.7> 2022-10-27T22:37:16.924522+02:00 Telcontar leafnode 7470 - - rereading /var/spool/news/leaf.node/groupinfo <7.7> 2022-10-27T22:37:16.936695+02:00 Telcontar leafnode 7470 - - <MODE READER <7.7> 2022-10-27T22:37:16.936781+02:00 Telcontar leafnode 7470 - - >200 Leafnode 1.11.11, pleased to meet you! <7.7> 2022-10-27T22:37:16.937236+02:00 Telcontar leafnode 7470 - - <GROUP rec.arts.tv <7.7> 2022-10-27T22:37:16.937264+02:00 Telcontar leafnode 7470 - - marked group rec.arts.tv interesting <7.7> 2022-10-27T22:37:16.937290+02:00 Telcontar leafnode 7470 - - markinterest: rec.arts.tv touched ctime <7.7> 2022-10-27T22:37:16.937315+02:00 Telcontar leafnode 7470 - - >211 219323 3 219325 rec.arts.tv group selected <7.7> 2022-10-27T22:37:16.948563+02:00 Telcontar leafnode 7470 - - <XOVER 3-219325 <7.7> 2022-10-27T22:37:16.948659+02:00 Telcontar leafnode 7470 - - markinterest: rec.arts.tv touched ctime <7.7> 2022-10-27T22:37:17.172800+02:00 Telcontar leafnode 7470 - - >224 Overview information for postings 3-219325: <7.7> 2022-10-27T22:37:26.915211+02:00 Telcontar leafnode 7480 - - config: debugmode is 1 <7.7> 2022-10-27T22:37:26.915325+02:00 Telcontar leafnode 7480 - - config: maxfetch is 5000 <7.7> 2022-10-27T22:37:26.915374+02:00 Telcontar leafnode 7480 - - config: maxage is 0 <7.7> 2022-10-27T22:37:26.915418+02:00 Telcontar leafnode 7480 - - config: postings have max. 500000 bytes <7.7> 2022-10-27T22:37:26.915460+02:00 Telcontar leafnode 7480 - - config: timeout_long is 100 days <7.7> 2022-10-27T22:37:26.915504+02:00 Telcontar leafnode 7480 - - config: timeout_fetchnews is 90 seconds <7.7> 2022-10-27T22:37:26.915548+02:00 Telcontar leafnode 7480 - - config: newsadmin is postmaster@telcontar.valinor <7.7> 2022-10-27T22:37:26.915592+02:00 Telcontar leafnode 7480 - - config: allow_8bit_headers is 1 <7.6> 2022-10-27T22:37:26.915634+02:00 Telcontar leafnode 7480 - - connect from localhost (::ffff:127.0.0.1) to localhost (::ffff:127.0.0.1) (my fqdn: Telcontar.valinor) <7.7> 2022-10-27T22:37:26.915680+02:00 Telcontar leafnode 7480 - - rereading /var/spool/news/leaf.node/groupinfo <7.7> 2022-10-27T22:37:26.935147+02:00 Telcontar leafnode 7480 - - <MODE READER <7.7> 2022-10-27T22:37:26.935269+02:00 Telcontar leafnode 7480 - - >200 Leafnode 1.11.11, pleased to meet you! <7.7> 2022-10-27T22:37:26.950569+02:00 Telcontar leafnode 7480 - - <GROUP rec.arts.tv <7.7> 2022-10-27T22:37:26.950667+02:00 Telcontar leafnode 7480 - - marked group rec.arts.tv interesting <7.7> 2022-10-27T22:37:26.950720+02:00 Telcontar leafnode 7480 - - markinterest: rec.arts.tv touched ctime <7.7> 2022-10-27T22:37:26.950759+02:00 Telcontar leafnode 7480 - - >211 219323 3 219325 rec.arts.tv group selected <7.7> 2022-10-27T22:37:26.970549+02:00 Telcontar leafnode 7480 - - <XOVER 3-219325 <7.7> 2022-10-27T22:37:26.970657+02:00 Telcontar leafnode 7480 - - markinterest: rec.arts.tv touched ctime <7.7> 2022-10-27T22:37:27.232311+02:00 Telcontar leafnode 7480 - - >224 Overview information for postings 3-219325: <7.7> 2022-10-27T22:38:04.043912+02:00 Telcontar leafnode 7480 - - <XHDR newsgroups 3-219325 <7.7> 2022-10-27T22:38:04.044290+02:00 Telcontar leafnode 7480 - - markinterest: rec.arts.tv touched ctime <7.7> 2022-10-27T22:38:04.044501+02:00 Telcontar leafnode 7480 - - >221 newsgroups: header (from article files) for postings 3-219325: <7.7> 2022-10-27T22:38:06.046798+02:00 Telcontar leafnode 7480 - - <XHDR user-agent 3-219325 <7.7> 2022-10-27T22:38:06.046885+02:00 Telcontar leafnode 7480 - - markinterest: rec.arts.tv touched ctime <7.7> 2022-10-27T22:38:06.046919+02:00 Telcontar leafnode 7480 - - >221 user-agent: header (from article files) for postings 3-219325: <7.7> 2022-10-27T22:38:13.509966+02:00 Telcontar leafnode 7480 - - Client timeout, disconnecting. <7.3> 2022-10-27T22:38:13.517137+02:00 Telcontar leafnode 7470 - - Cannot write to client: Broken pipe <7.7> 2022-10-27T22:38:13.517208+02:00 Telcontar leafnode 7470 - - Client timeout, disconnecting. - -- Cheers, Carlos E. R. (from openSUSE 15.3 x86_64 at Telcontar) -----BEGIN PGP SIGNATURE----- iHoEARECADoWIQQZEb51mJKK1KpcU/W1MxgcbY1H1QUCY1rtUxwccm9iaW4ubGlz dGFzQHRlbGVmb25pY2EubmV0AAoJELUzGBxtjUfVgM8AnjNjM6SJZxFkBWnjU60x oWbKKxDpAJ0Qp94sKUQgWbQbYvpX/ABZJYntbw== =NFmG -----END PGP SIGNATURE-----
On 10/27/22 15:42, Carlos E. R. wrote:
Repair folder sends TH into 100% CPU load and has to be killed again.
As can be seen in the log, TH fails to read from the server, it locks or crashes.
I'd search a bit on repair folder high cpu load and see if there is a mozilla article on how to handle it. If your usenet index is corrupt (sounds like it is), depending on the size, it can take a LONG time (multiple hours) to rebuild. Granted, it should not take 100% of the CPU, but some of the indexing code in tbird, especially with the circa 1980 "Mork" text database format for addressbook, etc.. When I've had to rebuild in the past, I usually let it bang away at 100% as long as I can stand it, then kill it as you did, restart and it tends to behave a bit better when not starting from scratch. There is an article on rebuilding the global indexes here. https://support.mozilla.org/en-US/kb/rebuilding-global-database -- David C. Rankin, J.D.,P.E.
On 2022-10-27 23:02, David C. Rankin wrote:
On 10/27/22 15:42, Carlos E. R. wrote:
Repair folder sends TH into 100% CPU load and has to be killed again.
As can be seen in the log, TH fails to read from the server, it locks or crashes.
I'd search a bit on repair folder high cpu load and see if there is a mozilla article on how to handle it.
The problem is known and was supposedly corrected with this update. So I was told here. NOT related to index building. I will bugzilla it. -- Cheers / Saludos, Carlos E. R. (from 15.3 x86_64 at Telcontar)
participants (2)
-
Carlos E. R. -
David C. Rankin