Dick van der Kaaden wrote:
> We have been using 1.4.1 for a while now on a redhat 7.0 with
> mysql 3.23.37, php 4.0.5 and apache 1.3.19. We frequently get a problem
> where the apache child process just segfaults. The mysql logs frequently
> shows : got error reading communication packet.
That could be cause or effect. When running apache stand alone (-X), is
this reproducable?
> In the apache error log there are sometimes messages like these:
>
> httpd: midgard.c:1191: mgd_release: Assertion `res && res->res' failed.
>
> httpd: format.c:46: mgd_clear_pool: Assertion `pool' failed.
Does anyone know a way to generate backtraces from within a program? It
would be most useful to have these asserts log a backtrace. In itself,
this message gives little clue on why res or pool is NULL.
> The first one occurs frequently the second one only incidentally.
> Further the logs also show many failed to connect to database errors
> (Midgard: switching to {db} failed). These result in article updates
> failing in the admin and users receiving 404 errors while visiting the
> sites. When they just press reload in their browser it works???
Alan posted a patch (now in CVS) that will probably fix this problem. I
don't know yet why the connection is being dropped but with the patch
Midgard will attempt a reconnect during the same request before giving up.
> What can we do about this? Also is there a new 1.4.2 release planned?
No date has been fixed, but we're talking about 1.4.2. There are a
number of bugs that I would like to see solved first.
Emile
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
