Start load
ssh test@kb-prod-udv-001 cd /home/test/bitarkiv-create-testdata-scripts/upload_test rm -r lib/* cp -r /home/test/release_software_dist/TEST11B/lib/* lib/. ssh kb-test-har-001 rm -rf UPLOAD-TEST ssh kb-test-har-001 mkdir UPLOAD-TEST scp -r * kb-test-har-001:/home/test/UPLOAD-TEST ssh kb-test-har-001 "cd UPLOAD-TEST;./batch_generate_stuff.sh 2" ssh kb-test-har-001 ". /etc/profile;cd UPLOAD-TEST;./batch_UPLOAD_1_run.sh > UPLOAD_1.out 2>&1 &" ssh kb-test-har-001 ". /etc/profile;cd UPLOAD-TEST;./batch_UPLOAD_2_run.sh > UPLOAD_2.out 2>&1 &" ssh kb-test-har-002 rm -rf UPLOAD-TEST ssh kb-test-har-002 mkdir UPLOAD-TEST scp -r * kb-test-har-002:/home/test/UPLOAD-TEST ssh kb-test-har-002 "cd UPLOAD-TEST;./batch_generate_stuff.sh 2" ssh kb-test-har-002 ". /etc/profile;cd UPLOAD-TEST;./batch_UPLOAD_1_run.sh > UPLOAD_1.out 2>&1 &" ssh kb-test-har-002 ". /etc/profile;cd UPLOAD-TEST;./batch_UPLOAD_2_run.sh > UPLOAD_2.out 2>&1 &" ssh netarkiv@sb-test-har-001 rm -rf UPLOAD-TEST ssh netarkiv@sb-test-har-001 mkdir UPLOAD-TEST scp -r * netarkiv@sb-test-har-001:/home/netarkiv/UPLOAD-TEST ssh netarkiv@sb-test-har-001 "cd UPLOAD-TEST;./batch_generate_stuff.sh 2" #Husk at rette /home/test til /home/netarkiv i batch_UPLOAD_?_run.sh via følgende kommando: ssh netarkiv@sb-test-har-001 ". /etc/profile;cd UPLOAD-TEST;sed -i 's/\/home\/test/\/home\/netarkiv/' batch_UPLOAD_*_run.sh" ssh netarkiv@sb-test-har-001 ". /etc/profile;cd UPLOAD-TEST;./batch_UPLOAD_1_run.sh > UPLOAD_1.out 2>&1 &" ssh netarkiv@sb-test-har-001 ". /etc/profile;cd UPLOAD-TEST;./batch_UPLOAD_2_run.sh > UPLOAD_2.out 2>&1 &" #start also 10 or more batchprogs (e.g. jove batchjobs) which use a lot of io and memory ( see privious step) #spam the archive with 200 concurrent batch jobs per sec. using following command: cd /home/test java -Xmx1536m -Ddk.netarkivet.settings.file=/home/test/TEST11B/conf/settings_ArcRepositoryApplication.xml -cp /home/test/TEST11B/lib/dk.netarkivet.archive.jar:. SendGetChecksumMessage KB 200
Control heap space og threads usage for the KB monitor
- login in to kb-test-adm-001.kb.dk as user test. Run
ps -ef | grep TEST11B
- Find the procesnumber for the KB monitor apps
- Start JConsole and select KB monitor procesnumber
- verify during some minutes, that heap space and number of threads are freed concurrently
Disable the network/switch/DC for some minutes and see that all batch processes reconnects and continue after restart
- login on to kb-test-adm-001 as root and stop the networkinterface by installing a cron-job that does this for you:
Install script restartNetworkWithWait.sh as root cronjob
#!/bin/bash # stopping network /etc/init.d/network stop # waiting 3 minutes /bin/sleep 3m # starting network /etc/init.d/network start
Go to the jconsole and reconnect again and verify that the processes continues go to the bitarchivemonitor in systemoverview and verify that there are still jobs running and uploads continues