[Date Prev] [Date Index] [Date Next] [Thread Prev] [Thread Index] [Thread Next]

Re: Config/operation question

Bryan Stansell bryan@conserver.com
Fri, 26 Mar 2004 05:01:30 -0800 (PST)


On Fri, Mar 26, 2004 at 07:29:12AM -0500, Pete Geenhuizen wrote:
>    Reintcheck as best I can tell sets up a delay or frequency for
>    conserver to check the connection to a particular host, and if the
>    connection is broken it will then re-connect, is that correct?

basically.  any and all consoles that are marked "down" (as in, not
connected) will attempt a connection every X minutes.

>    I've set the value to ten thinking it was seconds, but I believe
>    that its minutes according to the -O option, so I'll need to change
>    that. Either way I don't see that it has any affect,providing that
>    my assumption is correct.

yeah, it's really like the -O option, so minutes.  the conserver.cf
manpage is wrong...it'll say minutes instead of seconds in the next
release (which matches the -O docs).  sorry about that.

>    What I want to do/need is to be able to go to the terminal
>    concentrator and reset the port so that I can use the old fashioned
>    way of connecting to the console, and then when I'm done have
>    conserver pick up the console once again.
> 
>    Is reinitcheck the way to do that or is there another alternative, as
>    it stand now as soon as I issue a port reset conserver detects the
>    disconnect and immediately re-connects.

what you're seeing is expected...the "problem" is there are multiple
ways to tweak things and you've only pulled on lever #1.  here are the
other things to look at:

 1. autoreinit option
    this controls the automatic attempt to bring up the console (which
    is what you're seeing).  if the initial attempt doesn't succeed,
    it'll reattempt every minute thereafter.  this automatic behavior
    occurs only when a console connection is broken - not when a console
    has been taken down by a user (^Ecd).

 2. reinitcheck option
    this is a timer that goes off every X minutes and finds all the
    consoles that are down (from a broken connection or user command)
    and tries to bring it back up.

 3. reinitoncc option
    this will cause a connection attempt for a downed console (because
    of a broken connection or user command) when a client connects.
    this is off by default, so you may not need to worry about it.

so, if you enable reinitcheck and autoreinit (your case), you'll have
two behaviors happening, depending on how a console goes down.  if the
connection is broken, the autoreinit option will trigger an automatic
reconnection, with an attempt every minute as long as it fails.  the
reinitcheck timer will also be going off every X minutes and trying to
bring up the console (and any other consoles that are down).  if,
however, you connect to the console and do a '^Ecd', the autoreinit
option doesn't apply and you'll have anywhere from zero seconds to X
minutes before the reinitcheck timer goes off and attempts to bring the
console back up.  as long as you tie up the console in some other way,
conserver will reattempt that connection every X minutes, like you
expected.

so, as you can see, there are two key levers and one other interesting
one.  with that, you have 4 very interesting (and 4 more
semi-interesting) ways of configuring conserver regarding automatic
console initializations.  and i totally left out the 'ondemand'
option...but that doesn't really help you.  ;-)

hopefully with all those levers, you'll find a behavior you like.

Bryan