SRG sets redirecting

[SteveB1981]

I had posted this question some time ago but know one seem to have any idea. Furhter work has been done and the problem continues.

We have an SRG 50 running rls 2 connected to a CS1k 4.5.

We are basically seeing random phones reset either when idle or during a call. There is no drop in connection being seen between the SRG site and CS1K also to note it is not happening on all phones just random normally around 5 out of the 27 at a time over the space of 10 minutes.

This may happen once or twice in 24 hours.

I have turned on all logging possible and only log / alarm that is coming up is:

50075 DN:xxx Normal mode - set redirected to main office

Any help or comments much appreciated.

Thanks

Steve

 

[9201]

I have seem this problem in the pass, are all your phones static or dchp? If dhcp, i would check all switches and cable all the way to the DHCP server. no HUB in between.

Long time ago like 3yrs, i had the same issue, Nortel came out and found a HUB (no layer2) between the cisco switch and the DHPC server. We remove the HUB and all IP phones work fine.

 

[johnpoole]

couple of thoughts on the same line.. whats you tighest bandwidth between the server and the station? and are you running multicast, if dhcp how big is that broadcast domain? i've had to change from dhcp to static so that i could broadcast disable the network.. i could have left it dhcp if the customer would have enabled some broadcast relay on smaller domains.. have you put a data sniffer on that network

john poole
bellsouth business
columbia,sc

 

[SteveB1981]

Thanks guys for you comments so far...

I will put this into a visio diagram if that helps.

Basically this is across an MPLS link which has CoS for VOIP services applied to. I don't believe it is the WAN link.

The local network switch is a Nortel 5500 series on a flat network with around 30 users. The problem occurs outside of hours as well so I can't see anything effecting the QoS as there is literally nothing on the network apart from the phones outside of hours.

I'm not expert but can you explain how DHCP effect this is it during a renew request takes place and fails?

Please let me know any questions of set-up and I will gather, any suggestions on what to check on first?

Is it worth increasing the ip lease time and using static addressess or is it more broadcasting that could be the issue.

There are know hubs in the DHCP is directly hooked onto the the 5500 with the handsets.

If we were to put a packet sniffer on what should be looked for?

Many Thanks

Steve

 

[MagnaRGP]

sniff for traffic between one phone and the cs1k. i would do the same on the other side as well.

i think the heartbeat is getting lost through the mpls network.

 

[SteveB1981]

The phones in our main London office where the CS1K is located are fine no restarting just to confirm.

MPLS is a VPN based service so we can monitor this also it is not all phones it normally happens in batches of 5-7 out of the 30.

The problem is in one of our remote offices where the SRG is located.

What would the normal reason be why the heartbeat would be lost?

Is there anyway to increase the heartbeart delay? It also seems to happen mid call.

Many Thanks

Steve

 

[MagnaRGP]

i don't believe you can modify the heartbeat. my guess based on your description above is that the vpn service is timing out before the heartbeat from a phone.

the tunnel drops when not in use. then the heartbeat is sent from a phone and times out while the tunnel re-establishes. by this time the affected phone(s) have restarted and can reconnect immediately because the tunnel is up.

try setting up a continuous ping from london to the srg site from a pc and see if any of your phones reset at all while the ping is running. i'd wager they stay up.