Re: multilog error logging dnscache - supervise don't run

On 9/30/08, Paul Jarc <[email protected]> wrote:

> Your directory structure is not quite right.  If your dnscache
> configuration is in /etc/dnscache, for example, then /service should
> be a directory containing a "dnscache" symlink pointing to
> /etc/dnscache.
>
> I suspect your ran "ln -s /etc/dnscache /service" before you created
> the /service directory.  If that's true, then this will fix it:
> rm /service
> mkdir /service
> ln -s /etc/dnscache /service/

Unfortunately, that wasn't the cause - I did meanwhile install
daemontools on another PC and first configured /service, than
installed djbdns, with the same result - gdnscache is running, but
can't log - /service/log/main is empty. Only this time there is no
error in readproctitle regarding supervise.

Here is the output:

#ps auxww
root     24030  0.0  0.2   248   364 C0- S      3:42AM    0:01.37
readproctitle service errors: ...-data: not found\n./run[2]:
log-crucial-data: not found\n./run[2]:

#pstree

 |-+= 09393 root /bin/sh /command/svscanboot
 | |-+- 07304 root svscan /service
 | | |-+- 16308 root supervise log
 | | | \--- 31292 root (sh)
 | | |-+- 12517 root supervise dnscache
 | | | \--- 01447 gdnscache/usr/local/bin/dnscache /usr/local/bin/dnscache
 | | \-+- 25501 root supervise log
 | |   \--- 23918 gdnslog multilog t ./main
 | \--- 24030 root readproctitle service errors: ...-data: not found\n./run[2]:

Is it possible that the cause is that I created cruxlog, gdnscache and
gdnslog as "dumb" users, with shell /bin/false? Or an error in scripts
- did I make it wrong by creating them by "touch" instead by vi? Or
was it mistake to (in /service) mkdir ./log (with the dot) instead of
mkdir log (without a dot)?