[j-nsp] graceful failover and software upgrades

Jeff C. Strahl jeff.strahl at grandecom.com
Fri Mar 18 12:42:34 EST 2005

I opened a JTAC case on this during the period where vulnerabilities were
announced, and here was the verbiage I received from the JTAC:
Unfortunately, there is no specific documentation concerning upgrading a 2
RE router.  Like you said when we talked on the phone, you have to part and
piece the puzzle together.  I have tried to get you the information bit by
bit and these are as follows:

Before you upgrade, if you are running Graceful switchover, you must disable
it. (see the note section on the following page)


   the command to do this is:

   set chassis redundancy graceful-switchover disable 

1.   Enable Graceful restart to minimize/eliminate forwarding plane
disruption .  You also need to enable graceful restart on peer nodes for it
to buy you anything.
   the command to do this is:

   set routing-options graceful-restart enable restart-duration (seconds)

   for a overview of graceful restart see 


2. Upgrade backup RE by consoling in to it specifically
3. Manually switch (here is where graceful restart comes in) see the

4. Upgrade primary RE by consoling in to it specifically

5. If step 3 worked well, switch mastership back to primary RE
Please note Juniper does not claim support for hitless upgrades,
but this procedure has been known to result in pretty much that. All
bets are off if you are crossing major JUNOS software revisions. 


Hope this helps.

Jeff C. Strahl
Sr. IP Network Engineer
Packet Networks - Grande Communications
phone: 512.878.5465
cell: 830.708.6765
email: jeff.strahl at grandecom.com

-----Original Message-----
From: juniper-nsp-bounces at puck.nether.net
[mailto:juniper-nsp-bounces at puck.nether.net] On Behalf Of Roberts, Michael
Sent: Friday, March 18, 2005 11:24 AM
To: juniper-nsp at puck.nether.net
Subject: [j-nsp] graceful failover and software upgrades

Maybe I am missing something here.  We have several M20s that have redundant
REs.  Now that JNPR supports most of the major protocols and PICs (that we
have) in graceful fail-over, I want to enable this on our routers.  I have
started with enabling it on our lab chassis, and found that it works with
two routing engines running the same version of JUNOS.  What I do not
understand is the process for performing graceful failovers during a
software code upgrade.  If I have to disable graceful fail-over during a
software upgrade, then it does not do me much good unless I have a layer-3
failover that will route around the 'failed' node.  Maybe I am not
understanding the capabilities of the feature or not reading the docs
correctly?  Can anyone give me some insight on this?  Thx.



juniper-nsp mailing list juniper-nsp at puck.nether.net

More information about the juniper-nsp mailing list