A functional switch is not accessible via console port/SSH

Answered Question
Oct 23rd, 2007

Has anybody seen the following weird issue?

A switch 3750G can't be accessed via console port, the error message "

INSUFFICIENT MEMORY, TRY AGAIN LATER"

if accessed via SSH, it says "OPERATION TIMED OUT"

If accessed via CiscoWorks DFM, the alert showed "INSUFFICIENTFREEMEMORY"

Telnet is blocked. All switch functionalities are fine (including L3 fucntionality).

I don't think it is hardware issue. It looks like software issue(virus? IOS?) Anybody got any idea?

Your reply is appreciated.

Howard

I have this problem too.
0 votes
Correct Answer by Richard Burts about 9 years 3 months ago

Howard

This situation is not related to exec-timeout and configuring exec-timeout would not fix it - especially since you are not currently able to get access to the switch to get into configuration mode. While I believe that exec-timeout on vty ports to control remote access sessions is a very good idea, it is separate from your current issue.

As I commented in my reply in the WAN Routing and Switching forum, I suspect that your issue is related to a memory leak. I do not know a way to fix it other than a reload. As long as the switch continue to forward normally you have some flexibility about when to schedule the reload. I would be somewhat concerned that if the memory consumption gets worse that it might begin to impact normal forwarding and need an immediate reload.

HTH

Rick

  • 1
  • 2
  • 3
  • 4
  • 5
Overall Rating: 4 (1 ratings)
Loading.
glen.grant Wed, 10/24/2007 - 01:22

I think its reload time , have heard of something like this before . Also check to make sure all your lines have exec-timeouts on them , if this is not done then sessions get permanently tied up and will not free up after a specified time . On the console and vty lines add "exec-timeout 15" this will logout the session after 15 minutes of inactivity. you would have to look at the bug navigator for your ios version to see if any bugs exist, doubt its a virus.

master001 Wed, 10/24/2007 - 06:14

Thank you for your reply.

The situation is: I can't access the switch via console port since last week. Is it still exec-timeout issue?

Since the switch is not accessible now, how could I input command? Should I enter ROMMON mode?

Thanks again.

Howard

Correct Answer
Richard Burts Wed, 10/24/2007 - 09:07

Howard

This situation is not related to exec-timeout and configuring exec-timeout would not fix it - especially since you are not currently able to get access to the switch to get into configuration mode. While I believe that exec-timeout on vty ports to control remote access sessions is a very good idea, it is separate from your current issue.

As I commented in my reply in the WAN Routing and Switching forum, I suspect that your issue is related to a memory leak. I do not know a way to fix it other than a reload. As long as the switch continue to forward normally you have some flexibility about when to schedule the reload. I would be somewhat concerned that if the memory consumption gets worse that it might begin to impact normal forwarding and need an immediate reload.

HTH

Rick

master001 Wed, 10/24/2007 - 12:20

Thank you, Rick.

I will update you and rate it once I have a chance to test it.

Howard

master001 Thu, 10/25/2007 - 17:23

Hi Rick,

The switch rebooted itself, and it is accessible via SSH now.

It looks like memory leak.

Switch IOS version is: 12.2(25)SEB1

Anyway to prevent this switch from happening in the future? Anyway to monitor this switch memory status?

look forward to your reply. Thanks.

Howard

JBUIST Thu, 10/25/2007 - 17:37

Howard,

It sounds like you have LMS up and running. You can monitor the memory status using DFM and by tuning your thresholds, send alerts.

Jake

master001 Thu, 10/25/2007 - 17:53

Hi Jake,

Any CLI commands (including parameters) such as "show ..." to monitor switch memory?

Hi Rick,

This switch rebooted itself. Not rebooted as scheduled time. Is it related to memory leak?

If so, then back to previous post,what could fix memory leak issue? just reload ?

Howard

Richard Burts Thu, 10/25/2007 - 18:07

Howard

It is a bit hard to know whether the switch rebooted from the effects of a memory leak or from some other cause. It is possible (but not certain) that the output of show version or the output of show stack might provide some additional information. Or if you are sending syslog to a syslog server there might be helpful information in the syslog records.

But without knowing any more specifics, I would guess that the memory leak is the most likely cause of the reboot. In terms of how to fix the memory leak probably the "best" solution would be to upgrade the version of code. Hopefully versions that are more recent do not have this flaw. But pending a code upgrade you might try to monitor memory usage. If you have a feeling for how long the switch had been running it might give you an idea of how long you could go before the problem shows up again and you could schedule a reboot.

HTH

Rick

Actions

This Discussion