* Ralf Hildebrandt :
> * Ralf Hildebrandt :
>
> > Did that and recompiled, deactivated valgrind, let's see what happens.
>
> Right now I don't see any more libc related crashes.
Still no crashes. Your fix seems to be working
--
Ralf Hildebrandt (ralf.hildebra...@charite.de) snick...@c
* Ralf Hildebrandt :
> Did that and recompiled, deactivated valgrind, let's see what happens.
Right now I don't see any more libc related crashes.
--
Ralf Hildebrandt (ralf.hildebra...@charite.de) snick...@charite.de
Postfix - Einrichtung, Betrieb und Wartung Tel. +49 (0)30-450 5
* Timo Sirainen :
> I did find that pool_system_clean_realloc() didn't handle shrinking the
> memory area, fixed: http://hg.dovecot.org/dovecot-1.1/rev/17c73b14ed9d
I checked that out
> - default_pool = system_clean_pool;
> - data_stack_set_clean_after_pop(TRUE);
> + /*default
On Jan 15, 2009, at 7:28 PM, Ralf Hildebrandt wrote:
Well..
==10780== Invalid write of size 1
==10780==at 0x402499E: memcpy (mc_replace_strmem.c:402)
==10780==by 0x805B000: pool_system_clean_realloc (mempool-system-
clean.c:149)
==10780==by 0x804FC09: ssl_clean_realloc (ssl-proxy-
* Ralf Hildebrandt :
> > Are you also looking at what errors are in the valgrind log? It might
> > prevent the crashes. Does grep find any of:
> >
> > Invalid free() / delete / delete[]
> > Invalid read of size
> > Invalid write of size
> > bytes inside a block of size
>
> r...@postamt:/tmp# egr
* Timo Sirainen :
> On Wed, 2009-01-14 at 17:04 +0100, Ralf Hildebrandt wrote:
> > Jan 14 16:59:05 postamt dovecot: Dovecot v1.1.8 starting up
> > Jan 14 16:59:07 postamt dovecot: Fatal: imap-login: epoll_create(): Too
> > many open files
>
> Wonder if these patches fix this:
>
> http://hg.dovec
On Wed, 2009-01-14 at 17:04 +0100, Ralf Hildebrandt wrote:
> Jan 14 16:59:05 postamt dovecot: Dovecot v1.1.8 starting up
> Jan 14 16:59:07 postamt dovecot: Fatal: imap-login: epoll_create(): Too many
> open files
Wonder if these patches fix this:
http://hg.dovecot.org/dovecot-1.1/rev/66687cb6951
* Timo Sirainen :
> On Thu, 2009-01-15 at 16:42 +0100, Ralf Hildebrandt wrote:
> > * Timo Sirainen :
> >
> > > valgrind --log-file=/tmp/imap.%p could be helpful there then. But does
> > > it keep logging something? With me it logs one error when the process
> > > starts:
> >
> > A watched kettle
On Thu, 2009-01-15 at 16:42 +0100, Ralf Hildebrandt wrote:
> * Timo Sirainen :
>
> > valgrind --log-file=/tmp/imap.%p could be helpful there then. But does
> > it keep logging something? With me it logs one error when the process
> > starts:
>
> A watched kettle never boils. No incidents so far.
* Timo Sirainen :
> valgrind --log-file=/tmp/imap.%p could be helpful there then. But does
> it keep logging something? With me it logs one error when the process
> starts:
A watched kettle never boils. No incidents so far.
--
Ralf Hildebrandt (ralf.hildebra...@charite.de) snick...@cha
On Wed, 2009-01-14 at 15:31 -0500, Timo Sirainen wrote:
> Oh, right, libcrypto stuff isn't very valgrind-friendly. I think it's
> possible to disable these with:
>
>--suppressions= [default: $PREFIX/lib/valgrind/default.supp]
> Specifies an extra file from which to read desc
On Wed, 2009-01-14 at 21:26 +0100, Ralf Hildebrandt wrote:
> I get:
..
> ==9769== Conditional jump or move depends on uninitialised value(s)
> ==9769==at 0x40F09F7: BN_mod_inverse (in
> /usr/lib/i686/cmov/libcrypto.so.0.9.8)
> ==9769==by 0x40F48E0: BN_MONT_CTX_set (in
> /usr/lib/i686/cmov
* Timo Sirainen :
> Implemented: http://hg.dovecot.org/dovecot-1.1/rev/2146e1963b79
I rebuild dovecot just now with that
--
Ralf Hildebrandt (ralf.hildebra...@charite.de) snick...@charite.de
Postfix - Einrichtung, Betrieb und Wartung Tel. +49 (0)30-450 570-155
http://www.arschkre
* Timo Sirainen :
> valgrind --log-file=/tmp/imap.%p could be helpful there then.
Yes, done that now.
Better
> But does it keep logging something? With me it logs one error when the
> process
> starts:
>
> ==6036== Syscall param socketcall.sendmsg(msg.msg_iov[i]) points to
> uninitialised byte(s
On Wed, 2009-01-14 at 11:01 -0500, Timo Sirainen wrote:
> I was also thinking about adding some code that would attach IP address
> to the "killed with signal n" error (with
> login_process_per_connection=no it would be the last IP that started
> logging in).
Implemented: http://hg.dovecot.org/dov
On Wed, 2009-01-14 at 21:06 +0100, Ralf Hildebrandt wrote:
> * Timo Sirainen :
>
> > Ah, this is easy. Set login_process_size=0
> Ah, ok. Now it's very talkativ and clutters my log...
valgrind --log-file=/tmp/imap.%p could be helpful there then. But does
it keep logging something? With me it logs
* Timo Sirainen :
> Ah, this is easy. Set login_process_size=0
Ah, ok. Now it's very talkativ and clutters my log...
--
Ralf Hildebrandt (ralf.hildebra...@charite.de) snick...@charite.de
Postfix - Einrichtung, Betrieb und Wartung Tel. +49 (0)30-450 570-155
http://www.arschkrebs.de
On Wed, 2009-01-14 at 20:51 +0100, Ralf Hildebrandt wrote:
> * Timo Sirainen :
>
> > login_executable = /usr/bin/valgrind /usr/local/libexec/dovecot/imap-login
> >
> > It logs to Dovecot's error log (or use valgrind --log-file=/tmp/imap.%p).
>
> I remember last time:
>
> Jan 14 20:44:05 postam
* Timo Sirainen :
> login_executable = /usr/bin/valgrind /usr/local/libexec/dovecot/imap-login
>
> It logs to Dovecot's error log (or use valgrind --log-file=/tmp/imap.%p).
I remember last time:
Jan 14 20:44:05 postamt dovecot: imap-login: Valgrind's memory management:
out of memory:
Jan
On Wed, 2009-01-14 at 18:25 +0100, Ralf Hildebrandt wrote:
> * Timo Sirainen :
>
> > For non-SSL connections I think it would be usable (my machine handles
> > 600 logins/sec fine with valgrind), but I suppose with many SSL
> > connections it's too slow.
>
> How would I run imap-login from valgri
* Timo Sirainen :
> For non-SSL connections I think it would be usable (my machine handles
> 600 logins/sec fine with valgrind), but I suppose with many SSL
> connections it's too slow.
How would I run imap-login from valgrind?
--
Ralf Hildebrandt (ralf.hildebra...@charite.de) snick...
* Timo Sirainen :
> How soon after upgrading to 1.1.8 did this happen?
it always happened, even befor 1.1.8 :)
> There was one login process change done for 1.1.8, but it was supposed
> to fix a crash. And I don't really see how it could cause memory
> corruption. I guess you haven't seen these
On Wed, 2009-01-14 at 16:46 +0100, Ralf Hildebrandt wrote:
> * Timo Sirainen :
>
> > > Jan 14 08:42:52 postamt dovecot: imap-login: *** glibc detected ***
> > > imap-login: malloc(): memory corruption: 0x0844e778 ***
> > > Jan 14 08:42:52 postamt dovecot: child 15380 (login) killed with signal 6
* Timo Sirainen :
> > Jan 14 08:42:52 postamt dovecot: imap-login: *** glibc detected ***
> > imap-login: malloc(): memory corruption: 0x0844e778 ***
> > Jan 14 08:42:52 postamt dovecot: child 15380 (login) killed with signal 6
>
> The annoying thing is that these are pretty much impossible to f
On Wed, 2009-01-14 at 14:50 +0100, Ralf Hildebrandt wrote:
> >From my logs (dovecot-1.1.8)
>
> Jan 14 08:42:52 postamt dovecot: imap-login: *** glibc detected ***
> imap-login: malloc(): memory corruption: 0x0844e778 ***
> Jan 14 08:42:52 postamt dovecot: child 15380 (login) killed with signal 6
>From my logs (dovecot-1.1.8)
Jan 14 08:42:52 postamt dovecot: imap-login: *** glibc detected *** imap-login:
malloc(): memory corruption: 0x0844e778 ***
Jan 14 08:42:52 postamt dovecot: child 15380 (login) killed with signal 6
Jan 14 10:06:42 postamt dovecot: Panic: imap-login: file ioloop.c: l
26 matches
Mail list logo
