Multiple HP Procurve 5308xl / 2512 / 2650 .. etc.

[bruno maggi (o.tech)]

Hellooooo ....
I've a no so simple question related to a configuration done by the vendor that have made the deal on our customer. Basically we have 13 HP switch in a configuration as follows:

1) spanning-tree in RSTP-operation
2) 2 5308 in XRRP on ONE vlan
3) 10 switch connected by fiber with a redundant connection on the TWO core
4) 1 fiber that connect the two core switch
5) each vlan use the IP addess in the core switches (the same ip) as router
6) some HP server connected with a teaming configuration with one cable for each network card in each switch.

the HP partner have made a configuration without making any trunk or other change on ports. Just vlan assigned tagging the ports.

Now we are experiencing some trouble: port blocked/unblocked by STP, not just for the HP servers but also for some device with a single network adapter. I've done some change (enable / disable LACP etc.etc.) but no success ... and I have an IPSEC VPN with an online services that go down every 2-4 hours....

any suggestiones ?
thank you
brunom

[Evert Goor]

best is to change spanning tree priority's on the connection between the 2 5308 routers. Then you are allways sure that it will be the main root bridge for the spanningtree.
Look in the span stats if there is a lot of changes in the spanning tree. If yes look arond for faulty connections or unwanted swtiches in the network.
Keep at first the settings of the hp servers with trunking on auto. And have the settings on the switch with those ports clean.

If you still have trouble post it here

Good Luck

Evert

[bruno maggi (o.tech)]

Thank you for your reply.
I'm not a system network specialist on HP procurve so it's very difficult to translate you suggestion in a practical way .. but anyway you are suggesting "just" to change the the priority of all the link fiber that connect the switches as second path.

I've a default 128 that I've to change increasing the value (example 5000) for each fiber in the second switch ... (the second one redundant patch).

For the server I've followed the suggestion of the last document for the teaming nic (see the default chart) using the default / TLB suggested for the two switches case....

I've checked the last topology change is related to the last change (3 days) that I've done trying to solve the problem: the change was done adding a separated sw with ONE cascading port, with two port connected to the firewall.. (without success unfortunately... :))...

I try to configure a different prioriy on the links (but I've some concern related to the missing trunk configuration for the vlan and the generic configuration ...)

thank you I'll try to post the result of the tests asap ... any other suggestion is welcomed .....

grazie
brunom

[bruno maggi (o.tech)]

Two notes:

A- The policy for the downlink port for the redundant 5308 will be less than 128 ... But I think that the same operation need to be replicated on the other switches (link to the second 5308).... see the diagram. We have more or less all the vlan in all the switches
B- Do you think is better that I push a trunk on the two D1/D2 M1/M2 link in the slave side ? or in this case there is no need to use
thanks
bruno