HP2610 switcc keeps rebooting

[ZedViper]

Hi looking for a little help on trouble shooting a switch..a HP2610-48 has started rebooting itself unknown reasons. I have replaced this with an identical switch and a working config that is run throughout the buisness so I know the config is fine . I have then disabled all ports except for the two link ports 49 and 50 (the switch ran fine for over 30 minutes). Then renabled each port one by one and left a few minutes to see if the switch re-acts , at 20 it rebooted cpu usage at this time was normal, I disabled this port and after rebooting started again from 19 it rebooted again at 31 (did the same disabled port) carried on and it did it again . The strange thing is if I disabled the first 12 and ran from ports 13-24 it didnt trip.  I have replaced the switch with a 2530-24G to keep the offices in production for now and had no drops . So more curious for now on what could have been the cause and what could I have done differently to trouble shot this 

[Ivan_B]

Hi @ZedViper !

Is STP enabled?

[ZedViper]

Hi yes STP is enabled and loop guard protect is enabled as well. I feel like its a device on the network as it did the dsame with a replacement swithc but then under the same theory why is the 2530 working with no issue

[Ivan_B]

Just to confirm I understood your situation. You had a switch, its model is 2610-48. It started to reboot from time to time. Then you replaced this switch with identical one, again 2610-48 and uploaded the config from the previous device. The new switch also experience intermittent reboots and you think it has something to do with number of ports in the UP state. Then you replaced this second 2610-48 with 2530-24G and it works fine without any issues. Did I understand you correctly?

If so, it may be a software bug that is triggered by a traffic pattern. Of course I don't completely exclude the hardware issue, but two switches to fail at the same time... it's kind of unlikely.

Since STP and loop guard are enabled, we can exclude broadcast storm. Another confirmation it's not a storm is low CPU utilization before reboot is triggered.

I would do the following:

1. Upgrade to the latest s/w version R.11.123 and test
2. If the issue remains, contact the HPE Support. This model has Lifetime warranty.

[ZedViper]

Hi thats correct, both 2610-48 are in production in another part of the buisness which is non critical and are working fine. The switch had been upgraded to the latest version (sorry forgot to mention that)  what traffic pattern would cause such issues  and is there a way to trouble shoot it. logs were no good as the switch cleared these each time it re-booted.

CPU usage is only high when the switch starts up but thats to be expected and it goes up each time a port is enabled for a few seconds before settling.

Thanks 

[Ivan_B]

BTW, when did you perform the upgrade - before restarts started to happen or after that? No software is perfect, so if before swtiches were working fine, may be a downgrade to a reasonably fresh, but not the latest version could be tried as a troubleshooting step...

Normally nothing should cause a switch to reboot, but most often causes are due broadcast storms that lead to high CPU utilization. But even then it's not quite normal for a switch to reboot.

CPU utilization can be higher during the boot, because it loads the OS and all services. After that it should stay quite low, especially on such simple device that doesn't have any dynamic routing protocols etc...

"show tech all" is the command that generate an output that will be useful for HPE Support. Maybe it will contain some hints that can shed a light on what is going on. Some crash data can be there. Still, I'd suggest to open a case with support.

[ZedViper]

switch was upgraded after issue happened in an attempt to fix it.   Thank you for your help