Cannot restart Vertica 12.0 node when it is up
erkkik
Community Edition User ✭
I do not understand what is the recommended sequence of command line commands if i just want to restart vertica node what is up. I do not want to kill any process.
It will respond almost endlessly:
dbadmin@vertica-test1:~$ admintools -t restart_node -d CoopDWHTest -s v_coopdwhtest_node0001
Info: no password specified, using none
Nodes to be restarted but still not down{'v_coopdwhtest_node0001'}as of 2024-06-25 18:27:30. Next attempt after 5 secs
And if i want stop that node, it gives:
dbadmin@vertica-test1:~$ /opt/vertica/bin/admintools -t stop_node -s v_coopdwhtest_node0001
Sending signal 'TERM' to ['10.1.0.202']
Successfully sent signal 'TERM' to hosts ['10.1.0.202'].
Details:
Host: 10.1.0.202 - Success -
Checking for processes to be down
Skipping host 10.1.0.202 - no process signaled
No processes were signaled, skipping polling
Answers
as the messsage states , please wait for the admintools session to end.. it working on stopping the node.
"Nodes to be restarted but still not down{'v_coopdwhtest_node0001'}as of 2024-06-25 18:27:30. Next attempt after 5 secs"
command ends with:
Nodes to be restarted but still not down{'v_coopdwhtest_node0001'}as of 2024-06-25 18:37:21. Next attempt after 5 secs
Nodes to be restarted but still not down{'v_coopdwhtest_node0001'}as of 2024-06-25 18:37:26. Next attempt after 5 secs
Node v_coopdwhtest_node0001 is not down, can't restart
All nodes in the input are not down, can't restart
dbadmin@vertica-test1:~
@erkkik sorry for the confusion. it basically starts down node even though the name has restart. so you can use stop_node to stop the node and then start it using restart_node
https://docs.vertica.com/24.2.x/en/admin/using-admin-tools/admin-tools-reference/restarting-on-host/
As I wrote, stop_node command does not work.
dbadmin@vertica-test1:~$ dbadmin@vertica-test1:~$ /opt/vertica/bin/admintools -t stop_node -s v_coopdwhtest_node0001
Sending signal 'TERM' to ['10.1.0.202']
Successfully sent signal 'TERM' to hosts ['10.1.0.202'].
Details:
Host: 10.1.0.202 - Success -
Checking for processes to be down
Skipping host 10.1.0.202 - no process signaled
No processes were signaled, skipping polling
Node will stay in UP state and does not release memory.
Can anyone check, if it is problem with my instance or is it bug to stop node what is in UP state with command line command. I have Vertica 12.0 on Linux OS.