Still tracking down the Dspace/Tomcat problems.
Frank told me about the lsof command, which lists files opened by a process. trying that on badger gave no results. trying it on my Mac gave instant results. Hmmm…
I did a ps auxww | grep dsspace to see what was running, so I could get the right process name to feed to lsof. Found lots of dangling java processes, dating way back, plus some stuck cron jobs. tried to kill them, byt some wouldn’t die. Hmmm….
Then noticed that the latest active Dspace process was eating 99% of the CPU. Tried shutdown -r now. It shut down, but it didn’t restart. Frank had to go up and toggle the power.
After restart, needed to restart services that apparently still don’t restart manually:
- /etc/init.d/postgres start
- /etc/init.d/tomcat start
- /dspace/bin/start-handle-server
Odd — dspace continued to tell me that postgres wasn’t accepting connnections. Found postgres data in /var/lib/pgsql/data and /usr/local/pgsql/data, the latter being the real dspace data. Looks like an up2date in March created a new /etc/init.d/postgres , which I had to edit to reflect the old locations.
Finally, Dspace was back, and so far looks well-behaved.
Need to treat this server with kid gloves over the next couple of weeks until I see what is what. Meanwhile, Dspace 1.3.2 is out….
