RMI executor could not started

server
#1

Continuing the discussion from RMI executor could not be started:

Hi everyone,

I work on a vm Azure provided with Knime server. The vm is stopped each night and boot on the morning. And on the morning, when I try to launch a workflow, I have the error “RMI executor could not be started”.
I found that rebooting the tomcat/catalina server in sudo help to resolve the problem.
What do I need to do to not have this problem at the next reboot of the vm?

You will find in the attached file a part of the localhost.log file including the RMI error.

Thanks,

Lionel
error_RMI.txt (11.0 KB)

0 Likes

#2

Hi @lcollia,

Can you please describe how you shut down the server at night? And how you restart it the next day? The error points to some permission problems. Can you confirm that everything is shut down properly?

Apart from that: any specific reason why you stop the vm every night?

Cheers,
Roland

0 Likes

#3

Hi @RolandBurger,

Thanks for your reply.

Yes I was thinking the same, that is why I reboot tomcat in sudo.

The server is a vm on Azure (provided by MS and Knime). to the best of my knowledge, the shutdown/start procedure is directly controlled by the Azure portal. I do not have the details but I think that is ok.

Apart: yes a lot of reasons, the first one is the cost per hour of this preconfigure vm.

Regards,
Lionel

0 Likes

#4

Hi Lionel,

Just to confirm: When you start the VM via Azure portal, KNIME Server is started as well, but the executor has issues. You fix this by re-starting Tomcat, using sudo. Is that correct?

Also, are you using one of our pre-configured KNIME Server images on Azure, or did you install KNIME Server yourself?

Cheers,
Roland

0 Likes

#5

Hi Rolland,

You have all right.

I am using the preconfigured Knime Server image on Azure.

Cheers,

Lionel

0 Likes

#6

Hi Lionel,

From the looks of it, it seems like the permissions somehow got messed up at some point, which is why you need to use sudo now.

Just to be sure, when you restart tomcat, do you use sudo systemctl, or sudo startup.sh?

One thing you can try is doing a sudo chmod knime:knime /srv/ and /opt/, that could solve to issue.

Cheers,
Roland

0 Likes

#7

Hi Rolland,

I am using “sudo shutdown.sh” and “sudo startup.sh” cmd.

Thanks for your suggestion. I will not be able to test it before 1-3 weeks.

Best regards,

Lionel

0 Likes

#8

Hi Lionel,

Sounds good, let me know if this solves it once you try!

Cheers,
Roland

0 Likes

#9

Hi @RolandBurger,

I finally could try your solution without success.

I tried “sudo chown knime:knime” on /srv et /opt directories.

When the server restarted, the RMI was still down.

Cheers,

Lionel

0 Likes

#10

Hi Lionel,

Can you confirm that the error you were seeing after the changes was still the same? Do you see anything noteworthy in the logs?

Cheers,
Roland

0 Likes

#11

Hi Rolland,

Yes the error is the same.

Blockquote
In the /var/log/syslog file I have this :slight_smile:May 31 06:46:48 vmKnimeDv01 systemd[1]: apache-tomee.service: Control process exited, code=exited status=1
May 31 06:46:48 vmKnimeDv01 systemd[1]: apache-tomee.service: Unit entered failed state.
May 31 06:46:48 vmKnimeDv01 systemd[1]: apache-tomee.service: Failed with result ‘exit-code’.

But I do not know if it is related.

Cheers,
Lionel

0 Likes

#12

Hi @lcollia,

Sorry for the late reply, this got buried a bit. Are you still seeing the same issues?

Cheers,
Roland

0 Likes

#13

Hi @RolandBurger,

For various reasons I did not use the machine where I had the RMI error until here. I can confirm you that still I have the error. However we will have to build a new Knime server from the Azure store, so I hope I will not have the error on the new machine.

Cheers,
Lionel

0 Likes

#14

Hi Lionel,

Great, moving to a new instance will most likely take care of this as well!

Cheers,
Roland

0 Likes