Running the TCS on a robs1/tcs-test Tucson Machines

The test environment for Tucson is to use as the 64-bit TCS server and as the obs workstation (32-bit).

For the 2018A release, we got the latest mcspu running on robs1 as well (SVN version 835). This is not a simple task because typically, you don't need access to 32-bit and 64-bit TCS info on the same machine. But, when we run TCS GUIs from the 64-bit build and the mcspu simulator from the 32-bit build, we need both.

Had to start the servers on robs1 so that mcspu could use them. But, since we don't want TCS to think it's part of the cluster (don't want it to start subsystems there), we had to start the servers with the netconfig start passive

Note that tcs-test sometimes doesn't want to reboot nicely. There are disk mounting issues between tmpfs and nfs. The current configuration (Jan-2018) is not using the nfs disks for logs - so we have to clean up periodically. It's using local disks for /data - includes logs and telemetry.


Checked out the latest version into /home/ltscm/lbt32/mcspu/835

Stephen usually does this, so he manually creates bin and etc directories with the stuff used.
After the build, copied it to /lbt/mcspu/835

The gosim script used copies the config files to the local directory and then runs what SH calls mcsSimulator

What is the mcspu simulator doing with the ip address show command? Stephen said it was using it and so had to be root?

The TCSLOCALE has to be set to downtown or you get complaints about the IP address not being on the network.

Topic revision: r1 - 30 Jan 2018, KelleeSummers
This site is powered by FoswikiCopyright © by the contributing authors. All material on this collaboration platform is the property of the contributing authors.
Ideas, requests, problems regarding Foswiki? Send feedback