multilog error logging dnscache - supervise don't run

I have installed daemontools, ucspi-tcp and djbdns on OpenBSD 4.3.
stable, and set up external cache for home network according to the
instructions. Dnscache mostly works fine (though it can't resolve one
specific address), but I can't log it by multilog in
/service/log/main, although I've followed the instructions almost
strictly (details explained here
http://sites.google.com/site/sokotica/Home/djbdns-logging-with-multilog
at the site I created according to the instructions for getting help
from cr.yp.to).

Reported errors are:
# ps auxww
...
root      7958  0.0  0.2   196   380 ??  S      9:53AM    0:09.38
readproctitle service errors: ...rt supervise/run: file does not
exist\n./run[2]: log-crucial-data: not found\nsupervise: fatal: unable
to start supervise/run: file does not exist\
...
and
# pstree
...
 |-+= 00871 root /bin/sh /command/svscanboot
 | |-+- 04401 root svscan /service
 | | |-+- 06849 root supervise log
 | | | \--- 26227 gdnslog multilog t ./main
 | | |-+- 28791 root supervise dnscache
 | | | \--- 29133 gdnscache/usr/local/bin/dnscache /usr/local/bin/dnscache
 | | |-+- 02623 root supervise log
 | | | \--- 07467 root (sh)
 | | \-+- 05393 root supervise supervise
 | |   \--- 14743 root (supervise)
 | \--- 07958 root readproctitle service errors: ...rt supervise/run: file does

Scripts content:

# cat /service/log/run
#!/bin/sh
exec log-crucial-data
exec setuidgid cruxlog multilog t ./main

# cat /service/run
#!/bin/sh
echo starting
exec clockspeed
exec 2>&1

(I've been deleting and adding respectively exec 2>&1 in /service/run,
but the result is the same).

I think that the problem is either in ownership/permissions/groups
betweeen cruxlog and gdnslog, or I messed up something in scripts,
since I don't know how to write them.

Thanks in advance for any help.