[prev in list] [next in list] [prev in thread] [next in thread] 

List:       linux-ha-dev
Subject:    Re: [Linux-ha-dev] Re: [Linux-HA] Possible fix for double start problem.
From:       Alan Robertson <alanr () unix ! sh>
Date:       2003-09-22 17:57:55
[Download RAW message or body]

Luis Claudio R. Goncalves wrote:
> On Sun, Sep 21, 2003 at 12:33:09PM -0600, Alan Robertson wrote:
> | Lars Marowsky-Bree wrote:
> | >On 2003-09-19T21:12:06,
> | >   Alan Robertson <alanr@unix.sh> said:
> | >
> | >
> | >>The behavior I changed here is that if we try and start a resource group 
> | >>that we will check the status of each resource and start them if they 
> | >>don't report they're running.  In the past, it only looked at the first 
> | >>resource (commonly the IPaddr script), and then started them all if the 
> | >>first one wasn't running.  It does something similar for stopping 
> | >>resources.  For stopping resources I'm sure this is the right approach.  
> | >>For starting them, it might be an arguable choice...
> | >
> | >
> | >That is probably not an acceptable change for the STABLE series. People
> | >will be bitten by it if they use sloppily written custom resource
> | >scripts. Do you want to take bets? ;-)
> | >
> | >Most likely, the IPaddr is the first resource in most configurations, so
> | >that they likely never tested the other scripts... So for start, I'm not
> | >sure I agree with this change.
> | >
> | >I agree it's OK for stopping them; however, for stopping, I'd just
> | >iterate over all of them and stop them explicitly anyway, regardless of
> | >what the first resource says.
> 
> I agree with Lars and I'd also like to point that most of time people
> will be using LSB compliant initscripts instead of resource scripts.
> Looking for info on how to format status messages in a LSB fashion, I've
> found the following text:
> 
> "
> Error and status messages should be printed with the logging functions such
> as log_failure_msg and so on. Scripts may write to standard error or
> standard output, but implementations need not present text written to
> standard error/output to the user or do anything else with it.
> "
> http://www.linuxbase.org/spec/gLSB/gLSB/iniscrptact.html
> 
> This way, I think it can be a bit dangerous to rely on "OK" or "running"
> messages. Maybe we should stick to return codes and use these string as
> rule for resource scripts only.

I'm not inclined to fix that until we convert over to OCF-compliant resource 
scripts.

Heartbeat also runs on older non-LSB-compliant versions of Linux.  It's not 
Linux-HA resource scripts that need this.  It's needed for /etc/init.d 
scripts that aren't LSB-complaint...

-- 
     Alan Robertson <alanr@unix.sh>

"Openness is the foundation and preservative of friendship...  Let me claim 
from you at all times your undisguised opinions." - William Wilberforce

_______________________________________________________
Linux-HA-Dev: Linux-HA-Dev@lists.linux-ha.org
http://lists.linux-ha.org/mailman/listinfo/linux-ha-dev
Home Page: http://linux-ha.org/
[prev in list] [next in list] [prev in thread] [next in thread]