Showing results for 
Search instead for 
Do you mean 
Reply

Order to (re)start TC services

Whenever we need to reboot our TC servers (Windows patches, etc) I find that users are connectring their TC sessions only to the first of our 3 pool servers. (Note: JBOSS is running on pool server 1 as well.)
I restart the TC services on pool server 2 & 3, but this rarely changes anything.
If I reboot those servers, then all is well.
So, my basic question is: When I need restart the services (or reboot the servers themselves) what should be the order of events? (inlcuding the clean-up of various temp folders)

Gerry Champoux
Williams International
Walled Lake, MI
1 REPLY

Re: Order to (re)start TC services

Without knowing your exact landscape (servers and services) I can't give exact advice. Generally, shutdown all ancillary services first (subscription manager, FTS, cache, Dispatcher, etc.). Then shutdown the web tier followed by server manager(s). If this is truly a shutdown and not a deployment then stop FSC's too, slaves first then master. You can cleanup at this point before performing your backups, things like "clearlocks -assert_all_dead" are good options here. If this is a deployment then you have to leave the FSC master running but it won't stop you from cleaning up shared memory and other temp files.

Startup is the opposite, FSC's (master then slaves) first followed by server manager(s), web tier, then ancillary services. Give a 30 second pause after starting server manager to allow the tcserver processes to start and pause again (1 min?) after starting the web tier to allow jboss to initialize. Jboss and the server manager share (actually they each have their own and sync the tree) a treecache. Login issues can result if the services are started too quickly (before building the treecache is completed).

Randy Ellsworth, Teamcenter Architect, Applied CAx, LLC
NX 11.0.1.mp01 | SW 2016 | TcUA 11.2.3
Evaluating:AW 3.2