ISE2.2 console hangs

[bpaustia]

Recently installed ISE2.2 and seeing issue where a couple of times trying to login via console, SSH becomes un-responsive/hangs.

When I enter admin @ the console login prompt it hangs after hitting enter.  I never get prompted for a password?  Rebooting the box seems to clear the issue for a while, but eventually it hangs again.

Are there any specific ISE system logs, crashlogs, etc...  Recommended I can look at post-reboot to see if a process or something has stopped working that is causing console to become un-responsive?

-Bill

[hariholla]

Do you mean to say, when you login via Console, SSH service stops? Or are you always trying to access ISE CLI via SSH and seeing this issue? If you have console access and then SSH service stops, you can try disabling-enabling the SSH service with this global command on ISE CLI:

[no] service sshd enable

-Hari

[howon]

If this is a VM, ensure that the VM has enough RAM (8GB, 16GB preferred) dedicated for the system.

[switschel]

We are observing a similar problem with the evaluation ova. Whenever I work with the ISE, it hangs completely after less than 20 hours. The problem we see is as follows:

- ISE does not respond to any http(s) connects

- when connecting with SSH, we can enter username and password, but never get a CLI

- when connecting with the VM console, the ISE is completely unresponsive

rebooting the VM solves the problem for a while (in our case, less than 20hours). I took a show tech about 2 hours after the last reboot and see some swapping. 8 GB of memory already seem to be really tight; with one SSH and one WebSession, I see about 170MB of free DRAM, and 1 GB ob swap space usage.

With the current situation, I wonder if it is a good idea to use this version :-/. Any ideas where to look at are welcome. Could this be related to the way ressources are reserved within VMWare?

--suwi

[gbekmezi-DD]

I’ve noticed enough undesirable behaviors in the lab myself which lead me to the conclusion that waiting for more stability is a good idea before going to 2.2. Having said that, patch 1 was released yesterday…we’ll have to see how that goes. I didn’t notice anything in the release notes regarding stability which gives me hope, but that doesn’t necessarily mean those types of fixes weren’t included.

Having said that, I haven’t seen problems creeping up every 20 hours. They seem to be more random than that.

Regarding your specific question. Can you login from the console (server console/VMWare console) during the time you can’t login via SSH?

George

[switschel]

We actually do not work daily with the ISE, so the short period of stability may be a coincidence. The installation worked fine as long as it was not touched... which was about a week. However, when I configure, I won't be able to login the next day. The VM console does not answer in these cases, which makes data collection a little difficult. I suspect something to do with the VM setup.

As this will be a new setup - we only need TACACs. Anybody there who could say if there are differences regarding this between 2.0 and 2.1?

Susanne

[servicesecurity]

HI,

I am facing the same problem with OSE 2.2 patch 3.

I Gave the machine 32GB RAM, 8 CPU, 600GB Disk.

THe machine is freezing every 24 hours.

Is there any solution?????

REgards

[bravojared · Accepted solution]

If you are hanging every 24 hours and this is a VM, please ensure that you have disabled snapshots on the VM.  This will cause ISE to have issues that sound just like what you are seeing.  That is the likely cause here.

"The problem we see is as follows:

- ISE does not respond to any http(s) connects

- when connecting with SSH, we can enter username and password, but never get a CLI

- when connecting with the VM console, the ISE is completely unresponsive"

This is the exact experience when snapshots are happening.  Disable snapshots on those VMs, you do not need snapshots.  You have backups via other methods in ISE and snapshots are not supported.

[amunday]

We are experiencing this same behaviour on 2.2 patch 12. No shapshots are enabled across both HA nodes. We've escalated with TAC but is anyone else still seeing this behaviour?