[f-nsp] SX800 - hw-ip-next-hop limit reached

[Mike Schoenfeld]

I have a large VMware software testing environment that clones off large
amounts of machines for regression testing and I hit the hw-ip-next-hop
limit on my SX800 on Wednesday.  The SX800 provides routing for 3 subnets –
one a large /20 for the cloned testing machines.  We recently changed the
limit of “on” machines from 2000 to 3000 and the SX800 fell over with these
errors.


[image: Inline image 1]


I was able to change the default value of hw-ip-next-hop up to 3500 but
then the CPU started going crazy and sitting at 99%.  We set the VMware
limit back to 2000 machines and I changed the SX800 value back to default
of 2048 for hw-ip-next-hop and rebooted.  All is stable for now but I need
to find a solution to run more IP machines.


1.    I’m not really sure what hw-ip-next-hop is?  I assume this builds
some kind of forwarding table for each ARP entry?  Why is the IP-ARP limit
6000 if the box falls over after it has 2048 IPs in it?


2.    I have some new VDX gear in my office and I’m thinking of moving the
VMware environment over to a VDX 6710-54 so I can leave the SX800 for basic
server access.  All it needs to do is L3 routing for the testing
environment, no ACLs, no OSPF, no multicast, nothing but L3 routing for
3000 machines.  I can’t find any kind of default values for this thing.  Is
it going to fall over as well?  Has anyone seen any kind of hw-ip-next-hop
limit for the VDX series?

3. Not related, but VDX doesn’t support the IP Helper command “yet?”  Is
Brocade serious?


Thanks!

-Mike
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://puck.nether.net/pipermail/foundry-nsp/attachments/20140124/f1b606d0/attachment-0001.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: SX800_Error.jpg
Type: image/jpeg
Size: 182589 bytes
Desc: not available
URL: <https://puck.nether.net/pipermail/foundry-nsp/attachments/20140124/f1b606d0/attachment-0001.jpg>