hello
it works, but the logrotate is too long, and newer end in one day
so no segfault, but several logrotate process...
Still a good news.
Thanks
On Wed, 23 Jan 2019 22:42:53 +0100 =?UTF-8?Q?Bernhard_=c3=9cbelacker?=
wrote:
> Hello marc, dear Maintainer,
>
> a workaround could be to just
Hello marc, dear Maintainer,
a workaround could be to just to change the default stack
size in the script calling logrotate which is on stretch
maybe /etc/cron.daily/logrotate.
So maybe following line before calling
logrotate could be sufficient?
ulimit -s 16384
(I have not tested this.)
: Mardi 22 Janvier 2019 21:36:14
Objet: Bug#918106: logrotate: segfaults in rotateLogSet
Control: fixed 918106 logrotate/3.14.0-4
Control: tags 918106 = upstream fixed-upstream
Dear Maintainer, Hello Marc,
> (gdb) print log->numFiles
> $1 = 2122453
> stack size (kbytes,
Control: fixed 918106 logrotate/3.14.0-4
Control: tags 918106 = upstream fixed-upstream
Dear Maintainer, Hello Marc,
> (gdb) print log->numFiles
> $1 = 2122453
> stack size (kbytes, -s) 8192
That would match my assumption.
A maximum stack size of 8192 kb * 1024 /
r 6 in logrotate[562ea592f000+11000]
best regards
thanks
- Mail original -
De: "Bernhard Übelacker"
À: 918106-submit...@bugs.debian.org
Cc: 918...@bugs.debian.org
Envoyé: Lundi 21 Janvier 2019 21:57:16
Objet: Bug#918106: logrotate: segfaults in rotateLogSet
Hello Marc,
> but i
Hello Marc,
> but i don't see much more withour bt full, do i understand correctly ?
I just wanted to see the line from dmesg, where we can see which
address caused the error. And wanted to see all for the same crash.
And I had not realized that the callc instruction really have to
write its
00055e3239db8d2 : callq 0x55e3239d6ec0
0x55e3239db8d7 : nopw 0x0(%rax,%rax,1)
End of assembler dump.
- Mail original -
De: "Bernhard Übelacker"
À: 918106-submit...@bugs.debian.org, 918...@bugs.debian.org
Envoyé: Lundi 21 Janvier 2019 16:37:17
Objet: Bug#91810
Control: retitle 918106 logrotate: segfaults in rotateLogSet
Hello Marc,
I am sorry, but my advice to use 'bt full' makes
following commands to show the state of frame #1.
Therefore can you repeat the "coredumpctl gdb 754"
without the "bt full"?
Kind regards,
Bernhard
8 matches
Mail list logo
