SCN Constantly Going Down

[rval80]

I'm having an issue with the SCN that I set up. It's a 2 site SCN and the connection goes down every couple of days. Rebooting the system brings it back up, but after a while, I'm sure the customer is going to grow tired of rebooting the system constantly. does anyone else have this issue? Both systems are 8.0.43

 

[hairlessupportmonkey]

how are the switches linked? VPN, MPLS etc?

Id say its a network related issue, something like a firewall / router causing a shutdown of ports.

ACSS - SME
General Geek

 

[rval80]

It's a VPN link between the two sites. I think it's a network issue too, but the network guys keep telling me that if they can ping the two systems, then it's not their issue.

 

[tlpeter]

Pinging is not really a good tool for checking connectivity issues.
Yes it shows that it is up but that is about it.


BAZINGA!

I'm not insane, my mother had me tested!

 

[amriddle01]

Bollocks, put the two systems linked via a patch lead and show them it never goes down, put it on the network and it will, what does that prove?

 

[rval80]

So the IT guy had me do a telnet into the system, and he says that the port is open so he thinks the system needs to be tweaked. This is getting ridiculous.

 

[tlpeter]

What kind of routers are used?

BAZINGA!

I'm not insane, my mother had me tested!

 

[rval80]

Cisco 1921 Router

 

[amriddle01]

SCN is non configurable in that fashion, it just sends data on set ports to the specified IP address via the specified gateway, if there are issues with his traffic/link then it's the link at fault, just because they can't work out what it is/don't have the motivation to fix it doesn't change that fact. If the system was not configured correctly it would never work, dropping links after a time interval has firewall/router written all over it

 

[rval80]

Amriddle, I agree with you 100%. I've tried to tell them that if it was a system programming issue, it woudn't work ever. The only thing that's got me scratching my head is the fact that a reboot of the avaya brings it back up for a time before it drops again. Thats why these guys say it's the system.

 

[amriddle01]

That's because it starts a fresh connection to the other system opening the firewall port for return traffic, firewalls have timers that close this return path after certain intervals, I imagine this is the issue. Get a seperate system and build an SCN on the local LAN tprovitamins it, it may be the only way

 

[amriddle01]

Does rebooting the Cisco also bring it back up? Takes away that argument if it does

 

[amriddle01]

That should say "to prove this" not "tprovitamins" stupid bloody iPhone

 

[hairlessupportmonkey]

i suspect the cisco is has h323 and other SCN protocol fixups turned on. (depending on cisco os it might be inspect maps)

no fixup protocol h323 h225 1720
no fixup protocol h323 ras 1718-1719
no fixup protocol tftp 69

or

No Inspect H323 H225
No inspect H323 RAS

etc

ACSS - SME
General Geek

 

[rval80]

I haven't tried rebooting the Cisco. They rely heavily on their internet, so rebooting that is not a popular idea.

Hairless, I was told that there were no H.323 settings on the Cisco.

 

[amriddle01]

There nearly always is and they are on by default, if they have not turned them off then they are still on. Have they actually turned them off are they assuming that because they never turned them on then they must be off/not exist, I imagine not having a working telephony link is not conveniant either so a quick reboot, just one to narrow down the issue should be tolerated

 

[hairlessupportmonkey]

> They rely heavily on their internet, so rebooting that is not a popular idea

Also using the same broadband for their SCN and their day to day? Unless there is some pukka QoS then there is trouble right there.

ACSS - SME
General Geek

 

[rval80]

amriddle, I love the way you think because I feel the same way. I already told them that's what I am going to do and it has to be done for troubleshooting purposes.

Hairless, they only have a handful of phones, so I would think that wouldn't cause too much of an issue....I hope.

 

[rval80]

Actually hairless, I take that back. They have 2 connections here, but both route through the Cisco router. I believe one of them is specifically for voice traffic.

 

[IPOriface]

This is a nightmare problem which we've fought at a number of sites. It's always ended up being the router/firewall but proving it and more importantly, fixing it is a pain in the ARS.

One site we hit this at turned out to be a firewall NAT table issue. Flushing the firewall cache fixed the problem (until it reoccurred) - maybe you could get the Cisco guys to flush the IP tables on the routers and see if that brings SCN back up. Way less intrusive than rebooting the routers.

If this does fix it temporarily, check if the routers are covered by SmartNet, and then get the IT guys to contact Cisco for config assistance (good luck with the politics).