[j-nsp] M20 redundancy config?

[Neil Stirling]

Only one addition from me;

Use the feature 'keepalive-time' under [chassis redundancy].
Its turned off automatically by default, and when enabled it will allow the
backup RE to assume master control within x seconds if it detects loss of
keepalives from the master.

Default is 300 secs (5 mins), and the minimum is 2 secs.
******************************************************
Be aware though that pre-JUNOS 5.4 the minimum time is 300 seconds.
JUNOS 5.4 and after the minimum is 2 seconds.
******************************************************

>From the manual;

The range for keepalive-time is 2 through 10,000 seconds.

If you configure the keepalive time for 2 seconds, the sequence of events is
as follows:

After 2 seconds of keepalive loss, a message is logged. 
After 2 seconds of keepalive loss, the backup Routing Engine attempts to
assume mastership. An alarm is generated whenever the backup is active and
the display is updated with current status. 
Once the backup Routing Engine assumes mastership, it continues to function
as master even after the originally configured master Routing Engine has
successfully resumed operation. You must intervene to restore its previous
backup status. However, if at any time one of the Routing Engines is not
present, the other one becomes master automatically, regardless of how
redundancy is configured. 

OR

The range for keepalive-time is 300 through 10,000 seconds.

The sequence of events is as follows:

After 20 seconds of keepalive loss, a message is logged. 
After 300 seconds of keepalive loss (the default setting), the backup
Routing Engine attempts to assume mastership. An alarm is generated whenever
the backup is active and the display is updated with current status. 
Once the backup Routing Engine assumes mastership, it will continue to
function as master even after the originally configured master Routing
Engine has successfully resumed operation. Operator intervention is required
to restore its previous backup status. However, if at any time one of the
Routing Engines is not present, the other one becomes master automatically,
regardless of how redundancy is configured. 


HTH, Neil.

-----Original Message-----
From: billp [mailto:billp at wjp.net]
Sent: 23 May 2003 14:13
To: juniper-nsp at puck.nether.net
Subject: [j-nsp] M20 redundancy config?


After looking through all the docs and notes, I have come up with
the following config to enable reduncancy on an M20 (installed
dual RE and dual SSB):

chassis {
    redundancy {
        routing-engine 0 master;
        routing-engine 1 backup;
        ssb 0 preferred;
        failover on-loss-of-keepalives;
    }
}

I also use "commit sync" when committing configs, and try and
do a "request system snapshot" on both RE's individually, 
occasionally.

Question is- is this the proper configuration and operational
way to ensure redundancy in case of an RE or SSB failure?

Is there a better 'best practice'?

Any input appreciated.  Any other ideas/methods to increase
reliability/failover/etc would also be appreciated.

thanks
bill

_______________________________________________
juniper-nsp mailing list juniper-nsp at puck.nether.net
http://puck.nether.net/mailman/listinfo/juniper-nsp
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://puck.nether.net/pipermail/juniper-nsp/attachments/20030523/d4de9fc6/attachment.htm