Hi Dejan and others,

I do not have that hostcache file at all. You said I could ignore/delete it, 
but I don't have it to begin with (wince I can ignore/delete it it appears not 
having it is not such a big deal, but then why does it even exist, I wonder). 
Anyway, related to the nodes section you said  I should empty, that section 
seems to list the nodes in the cluster and that information seems to be 
cluster-specific, not host-specific (the same in both files  - one on each of 
my 2 nodes). If I just erase it, wouldn't I be killing useful information 
(which, again, seems to be the same on all nodes)?

But generally speaking yes, this is the kind of info I am looking for (ha.cf is 
indeed node-specific and thus I needs to handle it separately). Thanks for the 
advice. Please let me know if there's anything else you can think of. I need to 
operate at file level (I won't edit files in my procedure, I will handle each 
file as a whole).

Thanks!

----- Original Message ----
From: Dejan Muhamedagic <[EMAIL PROTECTED]>
To: General Linux-HA mailing list <[email protected]>
Sent: Saturday, December 15, 2007 10:06:41 AM
Subject: Re: [Linux-HA] cannot start heartbeat


Hi,

On Fri, Dec 14, 2007 at 03:53:06PM -0800, Radu Handorean wrote:
> Hi all,
> 
> When my computer boots heartbeat does not start automatically for
 some reason. I try to start it manually and here's what happens:
> 
> 16:16:42 [EMAIL PROTECTED] [~]> ps -A | grep heart
> 16:16:49 [EMAIL PROTECTED] [~]> /etc/init.d/heartbeat start
> Starting High-Availability serviceslogd is already running
> heartbeat[4750]: 2007/12/14_16:16:58 info: Version 2 support: true
> heartbeat[4750]: 2007/12/14_16:16:58 info: Enabling logging daemon 
> heartbeat[4750]: 2007/12/14_16:16:58 info: logfile and debug file are
 those specified in logd config file (default /etc/logd.cf)
> heartbeat[4750]: 2007/12/14_16:16:58 WARN: Core dumps could be lost
 if multiple dumps occur
> heartbeat[4750]: 2007/12/14_16:16:58 WARN: Consider setting
 /proc/sys/kernel/core_uses_pid (or equivalent) to 1 for maximum supportability
> heartbeat[4750]: 2007/12/14_16:16:59 info: **************************
> heartbeat[4750]: 2007/12/14_16:16:59 info: Configuration validated.
 Starting heartbeat 2.0.8
>                                                                      
 done
> 16:16:59 [EMAIL PROTECTED] [~]> ps -A | grep heart         
> 16:17:06 [EMAIL PROTECTED] [~]> /etc/init.d/heartbeat status
> Checking for High-Availability services                              
 dead
> 16:51:18 [EMAIL PROTECTED] [~]> 
> 
> (the time difference above is not how long it took to give me an
 answer, I just came back later with that command)
> 
> 
> Which is pretty much "nothing happens". I am running SUSE 10.2. What
 makes this a bit different than the usual setup is how I installed this
 machine. I made a tarball from another machine, tar-ing "everything".
 I'll skip the details (if there's anything you think it's important ask
 me and I'll explain it) but I used this tarball on another machine to
 set up the drive. Again, this is super high-level and superficial
 explanation but after having taken care of a whole bunch of detail, it
 works. I can boot my Linux, start X, do everything I want except start
 heartbeat.
> 
> Interestingly, if I use the tarball to restore the machine I created
 it from (unpack it on another partition and boot off of this new
 partition - again details related to the boot loader and other such are
 skipped here), it works just fine. It seems it's only when I use the tarball
 on another machine. To make everything else work I had to deal with
 files that contain host-specific info, such as mac addresses. I wonder if
 HA has any such files I need to handle separately.

/etc/ha.d/ha.cf lists hostname (uname -n) and, in case you run
crm (v2 config), /var/lib/heartbeat/crm/cib.xml and
/var/lib/heartbeat/crm/hostcache. The hostcache you can erase,
but the CIB you'll have to edit. The easiest is to empty the
nodes section.

Thanks,

Dejan

> 
> Any (other) ideas?
> 
> Thanks!
> 
> 
> 
>      
 
____________________________________________________________________________________
> Be a better friend, newshound, and 
> know-it-all with Yahoo! Mobile.  Try it now.
  http://mobile.yahoo.com/;_ylt=Ahu06i62sR8HDtDypao8Wcj9tAcJ 
> _______________________________________________
> Linux-HA mailing list
> [email protected]
> http://lists.linux-ha.org/mailman/listinfo/linux-ha
> See also: http://linux-ha.org/ReportingProblems

-- 
Dejan
_______________________________________________
Linux-HA mailing list
[email protected]
http://lists.linux-ha.org/mailman/listinfo/linux-ha
See also: http://linux-ha.org/ReportingProblems






      
____________________________________________________________________________________
Never miss a thing.  Make Yahoo your home page. 
http://www.yahoo.com/r/hs
_______________________________________________
Linux-HA mailing list
[email protected]
http://lists.linux-ha.org/mailman/listinfo/linux-ha
See also: http://linux-ha.org/ReportingProblems

Reply via email to