On Thu, Dec 21, 2000 at 09:22:40PM +0200, Sinisa Milivojevic wrote:
> Lutz Jaenicke writes:
> > On Thu, Dec 21, 2000 at 07:30:23PM +0200, Sinisa Milivojevic wrote:
> >
> > 1. It doesn't help (no surprise, as this would have meant that HP's compiler
> > and gcc having the same problem which is not very likely :-).
> > 2. I have to revise my statement above:
> > when receiving TERM, mysql reacts:
> > 001221 19:15:12 /usr/local/mysql1/libexec/mysqld: Normal shutdown
> > Just, the daemon does not die until recieving KILL...
>
> The above is definitely a problem with HP-UX signals. Try to set a
> breakpoint in our signal handler and see what happens.
Hi!
I have just found a bit of time to to trace through the code of mysqld.
My analysis finally ended in the function "close_connections()".
Here the following loop applies:
while (select_thread_in_use)
{
...
for (uint tmp=0 ; tmp < 10 ; tmp++)
{
error=pthread_cond_timedwait(&COND_thread_count,&LOCK_thread_count,
&abstime);
if (error != EINTR)
break;
}
...
}
This usage of pthread_cond_timedwait does not match the manual page
on HP-UX 10.20. On Linux it says: ...
The pthread_cond_timedwait function returns the following
error codes on error:
ETIMEDOUT
the condition variable was not signaled
until the timeout specified by abstime
EINTR pthread_cond_timedwait was interrupted by a
signal
On HP-UX 10.20 the manual page however states, that the "errno" variable
is set and the return value of pthread_cond_timedwait() can only be
0 or -1:
If the function fails, errno may be set to one of the following
values:
Return | Error | Description
_______|___________|_______________________________________
0 | | Successful completion.
-1 | [EINVAL] | The value specified by cond, mutex, or
| | abstime is invalid.
-1 | [EAGAIN] | The time specified by abstime expired.
-1 | [EDEADLK] | A deadlock condition is detected.
I don't have experience with thread programming (from comp.sys.hp.hpux
it seems that using 10.20 and doing thread programming is more or less
exclusive :-), so I am not sure whether this is the problem or only a
symptom thereoff...
Best regards,
Lutz
--
Lutz Jaenicke [EMAIL PROTECTED]
BTU Cottbus http://www.aet.TU-Cottbus.DE/personen/jaenicke/
Lehrstuhl Allgemeine Elektrotechnik Tel. +49 355 69-4129
Universitaetsplatz 3-4, D-03044 Cottbus Fax. +49 355 69-4153
---------------------------------------------------------------------
Before posting, please check:
http://www.mysql.com/manual.php (the manual)
http://lists.mysql.com/ (the list archive)
To request this thread, e-mail <[EMAIL PROTECTED]>
To unsubscribe, e-mail <[EMAIL PROTECTED]>
Trouble unsubscribing? Try: http://lists.mysql.com/php/unsubscribe.php
