Nortel stacks (BS 460) crash after power outtage.

[rizzah1977]

Hi,

We have many nortel stacks in our LAN/WAN (100/150 pcs). We got single units / stacks of 2 till 7). With many of our remote locations they keep regular power outtage tests. Alot of the times after such a test the flash got corrupt, or the stack didnt come back up again, or simply went doa on me. I have switches with the 3136 FW and 3162 versions, but that doesnt make a difference. Im wondering if anyone else have had simular problems with this.

thanx in advance.

 

[VOIPaintEASY]

Yes, we have the same issues though ours are 470/ 460 stacks. Nortel now says it is not a good idea to mix/ match 470/ 460's in a stack. Are you saying yours are just 460's and config get hanked up on power outages?

 

[HungryHouse]

I have always seen/heard that 470/460 stack combinations are indeed supported by Nortel...this is one of the "Allied" stack that Nortel refers to. The Key is to make sure that if you have 470-24 switches in the stack, it must be base.
I don't believe this is relevant to the original post, but rather to the reply by VOIPaintEasy.

If you know differently, feel free to indicate such.

Thanks, HH

 

[VOIPaintEASY]

We have stacks with 470's only that run fine. The mixed stacks have never been reliable long term. They will run, but do a Boos code upgrade campus wide on 10-15 stacks and 1-2 of them will have issues. A common issue on BOSS upgrades are corrupt passwords after the upgrade. Our latest bug is if a stack reloads from power issue the MLT and uplink are disabled. A bad position for a Network admin because all you can do at that point is take a laptop with console cable to the closet. That is assuming you are onsite, not out of town, etc. In any case we had a call with Nortel last Friday morning and they admitted there is widespread unexplained config corruption with mixed 460/ 470 stacks and they say we should break them up or replace them with their 5500 line. I am thinking something in a greenish blue color....

 

[rizzah1977]

VOIPaintEASY : My stacks are all 460's and i have also read that Nortel is not a big fan of mixing 460 and 470's. So i dont think im gonna try, since we allready have enough issue's with 460 stacks alone.

HungryHouse : I think you are correct, in a way that you can get a stack going with 460's and 470's if you make the 470 the base unit. However i didnt test this, and as i said above .. testing is it kinda useless since i wont mix them if the supplier isnt supporting it.

grtz
Peter

 

[HungryHouse]

One suggestion I would make is to use the Stack Monitor feature introduced in 3.1.8. This is extremely valuable for managing and troubleshooting stack disconnect issue and DMLT/MLT issues across a stack (power disconnects included).

Rel note of 3.1.8 gives good info on it.

-I hope this helps.

 

[VOIPaintEASY]

What issues do you have on 460 only stacks? Also, we had Nortel Pro Services design out network for the full CSE1000 VOIP rollout. They did put the 470's as base. We still have many bugs. If they are static and not changed or upgraded they are fine. But you evetually have to upgrade BOSS at some point. That is when the configs start getting tangled. As a note this is not apparent always right after the upgrade so it haunts you while you sleep... I think one good direction to go is a 470-POE only stack. They are 48 ports of POE. per 1U switch.

 

[rizzah1977]

My main problem is that when there are power outage tests that the 460 switches get corrupt. Flash memory becomes unreadable .. sometimes the switches die.

On other times when they power gets back .. and a stack gets online all at once, they dont get up oke, they dont know which is the base unit for example, in that case we have to turn off the stack again, and get them up one by one. This is ofcource when/if the stack gets back up normally

 

[VOIPaintEASY]

We have been advised from Nortel the best thing to do is default the config and re- enter the config on the new BOSS. I know this is lame advise from an enterprise level manufacturer. I am just passing along what we have experienced and been advised. I am about at the end of my Nortel data gear days. I can live with hard failures but software bugs and various housekeeping functions of a device like being able to store the NV RAM data reliably is a basic hardware manufacturing product function that has been a requirement for many years before this product line...

 

[rizzah1977]

VOIPaintEASY do you mean why re-enter the config the config file itself, or the Boss image ? cause thats what i have to do. I have to press ctrl-x during startup of the switch, then clear the logs and then upload and apply a new boss image on the switch before it can boot again.

 

[VOIPaintEASY]

No, what we ended up doing was defaulting the stack to factory default configs, then manually re- entering all config like VLANs, MLT, PVID, ports that were locked at 100 FDX, etc. Our latest issue was after a BOSS upgrade the stack would reload randomly. When the stack reset the MLT uplink ports would be disabled and also MLT would be disabled. That is not good as you then have to get to that closet with a laptop and a console cable to get the stack back to life. Also, as with other stacks after a BOSS upgrade this stack would corrupt the passwords. Luckily we could still access the stack via Device MGR and push a text config file to it with just passwords. This has happened on 3-4 other stacks. If you start getting corrupt telnet passwords that is a sign your config is corrupt and worse things are to follow. This is in our case always a side affect of a BOSS code upgrade. Nortel has no fix for it and in the end advises defaulting the stack.

 

[VOIPaintEASY]

Also, there is another know issue with Baystack switchs you should be aware of about the NMGT IP address going to slepp once a year.. I will start a new thread and post it.