• This site uses cookies. By continuing to use this site, you are agreeing to our use of cookies. Learn more.

Install Elastic Search on CentOS issue?

#1
I previously had Elastic search running perfectly however after a kernel panic on my hosts' node, the support has been unable to get Elastic Search installed/enabled again. I have pasted their responses below - if anyone can help me point them in the right direction for getting this running - that would be amazing!

I was able to start the elasticsearch init script without any errors. I didn't receive the sysctl error Chris received.

The application seemed to run for a few minutes, then stop.

It seems like perhaps the init script isn't able to set resource limits:
--
root@host [~]# sysctl vm.max_map_count
vm.max_map_count = 65530
++ MAX_MAP_COUNT=262144
root@host [~]# /etc/init.d/elasticsearch status 2>&1 | grep MAX_MAP
++ MAX_MAP_COUNT=262144
--

I will review the earlier ticket and update you as soon as I have more information.

Regards,
After having reviewed this issue, it appears as though the only plausible resolution to this issue is going to be to have either your developer look over this script or to contact the developer of the script about this issue and see what it is that they say needs to be done. If there are any server-side configurations that need to be made at the recommendation of the scripts developer or your own developer we will be happy to make those adjustments for you. But this is not simply just not something that we can do; we are not familiar with the script's code, itself, enough to make any alterations to it.
 

Xon

Well-known member
#3
Make sure you ensure the file ownerships of /var/lib/elasticsearch to be owned by the user for elasticsearch and not root.

ie, assuming elasticsearch is the right user:
Code:
sudo chown -R elasticsearch:elasticsearch /var/lib/elasticsearch
 

MattW

Well-known member
#6
Appears to be this issue. What are the possible solutions, since we are on a VPS and kernel cannot be changed?
At the minute, nothing. There is no fix available. If it was working before, then your host must have updated the kernel during the reload to bring the new one in.
 

MattW

Well-known member
#8
Then you need to wait for the host to decide if they are going to roll out that kernel to their nodes. One of my customers is using GoDaddy (I know, I know!), and they don't apply EVERY new kernel update and review and test, so could be a while off yet.