| Summary: | app-office/akonadi-server-1.1.1 crash | ||
|---|---|---|---|
| Product: | Gentoo Linux | Reporter: | renato gallo <renatogallo> |
| Component: | [OLD] KDE | Assignee: | Gentoo KDE team <kde> |
| Status: | RESOLVED FIXED | ||
| Severity: | major | CC: | cuciferus, dschridde+gentoobugs, genzilla, lfranchi, renatogallo, scientica, tl |
| Priority: | High | ||
| Version: | unspecified | ||
| Hardware: | AMD64 | ||
| OS: | Linux | ||
| Whiteboard: | |||
| Package list: | Runtime testing required: | --- | |
| Attachments: | akonadi debug crash log | ||
|
Description
renato gallo
2009-04-26 09:24:06 UTC
Created attachment 189472 [details]
akonadi debug crash log
attached akonadi debug crash log
I would particularly turn your attention on this line of the crash log [akonadiserver] akonadiserver: libmysql.c:4301: setup_one_fetch_function: Asserzione `param->buffer_length != 0' fallita. was this complete enough or you need more datas ? (In reply to comment #2) > I would particularly turn your attention on this line of the crash log > > [akonadiserver] akonadiserver: libmysql.c:4301: setup_one_fetch_function: > Asserzione `param->buffer_length != 0' fallita. > "[ 0: akonadiserver(_Z11akBacktracev+0x39) [0x40b353] 1: akonadiserver [0x40b877] 2: /lib/libc.so.6 [0x7f0e30705270] 3: /lib/libc.so.6(gsignal+0x35) [0x7f0e307051e5] 4: /lib/libc.so.6(abort+0x183) [0x7f0e30706703] 5: /lib/libc.so.6(__assert_fail+0xe9) [0x7f0e306fe229] 6: /usr/lib/libmysqlclient_r.so.15(set_stmt_errmsg+0) [0x7f0e2a54b1db] 7: /usr/lib/libmysqlclient_r.so.15(mysql_stmt_bind_result+0x151) [0x7f0e2a54c50d] 8: /usr/lib64/qt4/plugins/sqldrivers/libqsqlmysql.so [0x7f0e2a8b94ae] 9: /usr/lib64/qt4/plugins/sqldrivers/libqsqlmysql.so [0x7f0e2a8b8a4e] 10: /usr/lib64/qt4/libQtSql.so.4(_ZN9QSqlQuery4execERK7QString+0x126) [0x7f0e323f69a8] 11: /usr/lib64/libakonadiprivate.so.1(_ZN13DbInitializer10checkTableERK11QDomElement+0x2ec6) [0x7f0e329e05fc] 12: /usr/lib64/libakonadiprivate.so.1(_ZN13DbInitializer3runEv+0x51b) [0x7f0e329e3cd9] 13: /usr/lib64/libakonadiprivate.so.1(_ZN7Akonadi9DataStore4initEv+0x57) overlay ver crashes too [0x7f0e329d5237] 14: /usr/lib64/libakonadiprivate.so.1(_ZN7Akonadi13AkonadiServerC1EP7QObject+0x57f) [0x7f0e329676bd] 15: /usr/lib64/libakonadiprivate.so.1(_ZN7Akonadi13AkonadiServer8instanceEv+0x35) [0x7f0e32968473] 16: akonadiserver(main+0x2dd) [0x4067c4] 17: /lib/libc.so.6(__libc_start_main+0xe6) [0x7f0e306f15c6] 18: akonadiserver [0x406419] ] I can confirm, using all the kde-live packages. same crash. any news on this ? (In reply to comment #5) > I can confirm, using all the kde-live packages. same crash. > news ? I'm having the exact same problem on kde-4.3 with akondi-server-1.2, mysql-5.0.83 and qt-sql-4.5.2 Any news on this? Without working akonadi lot of stuff from kde is kind of useless. I am waiting for answers like you (In reply to comment #8) > I'm having the exact same problem on kde-4.3 with > akondi-server-1.2, mysql-5.0.83 and qt-sql-4.5.2 > > Any news on this? Without working akonadi lot of > stuff from kde is kind of useless. > Could this possibly be related to this[1] old (feb. 2008) bug in MySQL? (I'm just guessing based on the error message.) That bug appears to have been closed because it wasn't reproducable anymore. [1] http://bugs.mysql.com/bug.php?id=25251 my 'versions': AMD64 KDE 4.3 akonadi-4.3.0 akonadi-server-1.2.0 mysql-5.0.76-r1 qt-sql-4.5.1 I'm seeing the same bt as Comment #4 (different addresses though) bug of fame in the Akonadi Userbase FAQ: http://userbase.kde.org/Akonadi#The_so-called_.22Gentoo-Assert.22 Can confirm the issue with dev-db/mysql-5.0.76-r1 and dev-db/mysql-5.0.83, both times using x11-libs/qt-sql-4.5.2. Log from 5.0.76-r1: [akonadiserver] Database "akonadi" opened using driver "QMYSQL" [akonadiserver] DbInitializer::run() [akonadiserver] checking table "SchemaVersionTable" [akonadiserver] checking table "ResourceTable" [akonadiserver] checking table "CollectionTable" [akonadiserver] akonadiserver: libmysql.c:4301: setup_one_fetch_function: Assertion `param->buffer_length != 0' failed. [akonadiserver] "[ [akonadiserver] 0: akonadiserver(_Z11akBacktracev+0x39) [0x40ab29] [akonadiserver] 1: akonadiserver [0x40b05a] [akonadiserver] 2: /lib/libc.so.6 [0x7fb5b2acc460] [akonadiserver] 3: /lib/libc.so.6(gsignal+0x35) [0x7fb5b2acc3d5] [akonadiserver] 4: /lib/libc.so.6(abort+0x183) [0x7fb5b2acdfe3] [akonadiserver] 5: /lib/libc.so.6(__assert_fail+0xe9) [0x7fb5b2ac5429] [akonadiserver] 6: /usr/lib/libmysqlclient_r.so.15 [0x7fb5b060652c] [akonadiserver] 7: /usr/lib/libmysqlclient_r.so.15(mysql_stmt_bind_result+0x14c) [0x7fb5b0607b1c] [akonadiserver] 8: /usr/lib64/qt4/plugins/sqldrivers/libqsqlmysql.so [0x7fb5b0979427] [akonadiserver] 9: /usr/lib64/qt4/plugins/sqldrivers/libqsqlmysql.so [0x7fb5b0978329] [akonadiserver] 10: /usr/lib64/qt4/libQtSql.so.4(_ZN9QSqlQuery4execERK7QString+0x126) [0x7fb5b25804d6] [akonadiserver] 11: /usr/lib64/libakonadiprivate.so.1(_ZN13DbInitializer10checkTableERK11QDomElement+0x3001) [0x7fb5b40dffa1] [akonadiserver] 12: /usr/lib64/libakonadiprivate.so.1(_ZN13DbInitializer3runEv+0x514) [0x7fb5b40e3cd4] [akonadiserver] 13: /usr/lib64/libakonadiprivate.so.1(_ZN7Akonadi9DataStore4initEv+0x6b) [0x7fb5b40d12cb] [akonadiserver] 14: /usr/lib64/libakonadiprivate.so.1(_ZN7Akonadi13AkonadiServerC1EP7QObject+0x585) [0x7fb5b405fb15] [akonadiserver] 15: /usr/lib64/libakonadiprivate.so.1(_ZN7Akonadi13AkonadiServer8instanceEv+0x4a) [0x7fb5b406091a] [akonadiserver] 16: akonadiserver(main+0x3a2) [0x405df2] [akonadiserver] 17: /lib/libc.so.6(__libc_start_main+0xe6) [0x7fb5b2ab79f6] [akonadiserver] 18: akonadiserver [0x405959] [akonadiserver] ] [akonadiserver] " ProcessControl: Application 'akonadiserver' returned with exit code 255 (Unknown error) "akonadiserver" crashed too often and will not be restarted! The issue might vanish when compiling dev-db/mysql with USE=-debug. (But since other errors appear in that case, I can not be sure.) @ Comment #13 : I tried re-emerging mysql with USE=-debug and ran "akonadictl start". It seems to be running ok now, it didn't crash. Got some error about some resource not being migrated (but most of them appears to have been). Though, I'm suspecting that -debug simply let's the error slip though (not doing abny code for the assert macro). That or maybe that assertion shouldn't be there at all (but then other distros should've hit the bug too :/)? Btw, the ebuild warns about -fPIC, I tried "CFLAGS="-O2" CXXFLAGS="-O2" USE=-debug emerge mysql" but the fPIC warning remained so I just re-emerged with "USE=-debug emerge mysql" (thus with optimizations). USE=debug on mysql enables a LOT of assertion hard checks (with exit on failure). Why are you using it? (In reply to comment #15) > USE=debug on mysql enables a LOT of assertion hard checks (with exit on > failure). Why are you using it? I was debugging vanishing records in the Akonadi database. Assuming akonadi isn't using MySQL's embedded mode, you'd be better off using the querylogs than USE=debug. I'll put larger warnings around USE=debug for MySQL, but you shouldn't be using it in this case. Removing mysql team now. The Akonadi database strongly misbehaves (the abovementioned vanishing records). Binary logs show succeeding insertion of huge numbers where there should be strings, and huge numbers where there should be small numbers. The reason seems to be indeed Gentoo's MySQL (-patches?), as it does not happen on other distributions or when using SQLite as backend. had the same problem on slackware (In reply to comment #18) > The Akonadi database strongly misbehaves (the abovementioned vanishing > records). Binary logs show succeeding insertion of huge numbers where there > should be strings, and huge numbers where there should be small numbers. The > reason seems to be indeed Gentoo's MySQL (-patches?), as it does not happen on > other distributions or when using SQLite as backend. > also on other distros http://chakra-project.org/bbs/viewtopic.php?id=229 (In reply to comment #18) > The Akonadi database strongly misbehaves (the abovementioned vanishing > records). Binary logs show succeeding insertion of huge numbers where there > should be strings, and huge numbers where there should be small numbers. The > reason seems to be indeed Gentoo's MySQL (-patches?), as it does not happen on > other distributions or when using SQLite as backend. > any news on this? what happens with 1.2.1 or 1.3.1 ??? Still the same with app-office/akonadi-server-1.3.1 kde-base/akonadi-4.4.2 x11-libs/qt-sql-4.6.2 dev-db/mysql-5.0.90-r2 Removing the debug use flag from mysql still fixes it. still there at kde 4.4 - and probably a lot more relevant since it seems that akonadi is now required for kde4.4 (at least if you want to read your email) As per http://userbase.kde.org/Akonadi_4.4/Troubleshooting#The_so-called_.22Gentoo-Assert.22 have you tried rebuilding x11-libs/qt-sql since you last rebuild mysql. Seems there can be protocol mismatches between the client and server if you don't. No news for many months, let's suppose it is fixed. Reopen if it persists |
