ACS Engine Hanging / Replication Problems

[omega7]

I have two ACS 1112 Appliances running the latest software (Release 4.0(1) Build 42). Each appliance seems to run fine on its own. However, after setting up and successfully performing replication, the second ACS will not fully reboot. It says CSAuth did not start. 'show' usually shows the cpu at 100% with the services in various states of stopped, stopping, or starting. The web interface is unavailable. Another thing I have noticed that I think may have something to do with it is the status of the remote agents in the network device table. After replication, (and before rebooting) I can click on one successfully on the original machine, but when I attempt to click on one on the second appliance, I get a 404 browser error, and my ACS session is closed. I have to log back in to do anything else. Right now, I am rebuilding the second appliance from the cd (for the 15th time) to attempt replication with no remote agents defined to make narrow down the problem. Also note that if I manually add a remote agent on the second appliance, I am able to get to its properties with not problems. I am only not able to get to replicated entries' properties. Thanks in advance for any help.

[omega7]

Well, forget about the remote agents. The primary appliance has a very basic config. The only things in the network device table are itself and the other ACS. They each have the correct settings and the same key. The backup ACS has no configuration settings, except the ACS settings in the network device table and the appropriate replication settings. After a successful replication from primary to backup, and a reboot of the backup--it will not start back up. The CPU is at 100% and the services look like this:

CSAdmin stopped

CSAuth starting

CSDbSync starting

CSLog stopping

CSMon starting

CSRadius starting

CSTacacs starting

CSAgent running

thanks.

[darpotter]

Sounds like a bug in ACS.

Something in you config is causing CSAuth on the slave to blow up. This shouldnt happen and you should open a TAc case and make sure its escalated to the DEV team.

[rchavers]

I have just discovered the same problem. I set up 2 ACS appliances. I was able to replicate from the primary to the secondary with no problems until I rebooted the secondary ACS. Now I cannot connect using http. When I connect to the console and do a show command I see all of the services "starting" CSAdmin is stopped and the CPU is running 100%.

Looks like a call to TAC

[omega7]

This problem went away after I upgraded to 4.1. If you do upgrade, be sure you apply 4.1.1.23. It includes: CSCsg97429 - TACACS+ Command Accounting does not work in ACS 4.1(1) Build 23, among other things.