httpd-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Chris Darroch <chr...@pearsoncmg.com>
Subject Re: [PATCH]: Call dbd_setup() for all virtual hosts or create mutex in mod_dbd.c
Date Fri, 11 May 2007 03:58:11 GMT
Nick Kew wrote:

> I was wondering about that, but reluctant to propose a backport
> from trunk without doing some more research.  If you want to make it
> a backport proposal, I'll try and get my brain around it (and one or
> two related issues) in the morning.

   The main thing I'd point to is this long explanation of what the
issues I found were:

http://marc.info/?l=apache-httpd-dev&m=116742014418304&w=2

   If you follow the links at the bottom of this page, it maps my
original patches to what eventually got committed to trunk, along
with the comments in SVN:

http://people.apache.org/~chrisd/patches/mod_dbd_pools_groups/


> Anyone here using trunk mod_dbd operationally and have anything
> to report?
 
   We just drop the mod_dbd.c from trunk into 2.2.x and it works fine.
I did some really extensive testing of it beforehand, checking
all the various combinations of options, and I think it's pretty solid.

   IIRC, r491884 contains a number of fixes for old problems with
somewhat unusual combinations of things, like the use of ap_dbd_acquire()
on a non-threaded platform.  I tried to check all the combinations of
the following:

- with and without threads
- with and without persistent connections
- ap_dbd_[c]acquire(), with release during request or connection cleanup
- ap_dbd_open(), with explicit release using ap_dbd_close()
- failure during setup when creating persistent connection pools

   That's off the top of my head, so I may have missed a few things;
it's been a while since I thought about it.

   I'll try to start making up backports but it won't be tomorrow,
for sure.  :-/

Chris.

-- 
GPG Key ID: 366A375B
GPG Key Fingerprint: 485E 5041 17E1 E2BB C263  E4DE C8E3 FA36 366A 375B


Mime
View raw message