bryan at conserver
May 11, 2006, 7:06 AM
Post #2 of 2
(744 views)
Permalink
|
well, a very simple, quick test on my end worked. are you seeing any
parse errors when conserver loads the config? and is the initdelay
getting into the appropriate config{} block? running conserver with the
-S flag to test the config would be interesting (so errors can be easily
seen).
looking back through the users mailing list, you talked about using ssh
to connect to the cyclades. is that still true? if so, are you setting
the 'host' attribute to anything, as that's what the initdelay option
needs to group initializations (set it to any string if it isn't used).
Bryan
On Wed, May 10, 2006 at 03:14:22PM -0700, Phil Dibowitz wrote:
> So some of you may remember the issues I've had with
> conserver->cyclades. As a quick summary we have stupid firewalls in
> between that arbitrarily kill connections and the TS-series cyclades
> seem to get very very confused by this.
>
> After a lot of debugging with Cyclades Support, they sent me out an
> ACS-series to test with, which is the replacement for the TS series, but
> with almost entirely new code (2.6 kernel, etc.). Initial testing shows
> that the old bug seems to be gone, but I am having a few strangenesses.
>
> However, in the process, I think I found a bug. When I set "initdelay
> 60", it still seems to init all consoles on a given cyclade at the same
> time... this is contrary to the desired effect....
>
> --
> Phil Dibowitz
> P: 310-360-2330 C: 213-923-5115
> Unix Admin, Ticketmaster.com
>
_______________________________________________
users mailing list
users[at]conserver.com
https://www.conserver.com/mailman/listinfo/users
|
signature.asc
