Elastic search won't start on 22.1

[Giant850]

Just upgraded to Opnsense 22.1, then upgraded Zenarmor. Now Elasticsearch won't start. Packet engine and cloud agent are running. Ideas?

[mb]

Hi @Giant850,

Can you run the following two commands and paste the results here:

Code: [Select]

pkg info | grep elasticsearch | cut -d" " -f1 | xargs pkg info

Code: [Select]

service elasticsearch status

[RamSense]

I had the same.
Under Zenarmor - status -> Engine Version - I hit check updates and also App & Rules DB Version - I hit check updates.
after that I got an green notification on top from zenarmor about "noticing OS change" or something and did some reinstall of elastic search, after the notification the system came back online.

[mb]

@RamSense,

Yes, this is how it's supposed to happen.

After you switch to 22.1, zenarmor should detect the ABI change and re-install database with the correct ABI.

[ArminF]

Morning,

with my update to 22.1 the elasticsearch does not start.
"Error occured during elasticsearch start process"

installed and reinstalled
elasticsearch5   5.6.16_8   72.9MiB   SunnyValley   APACHE20
os-sensei   1.10.1   120MiB   SunnyValley   BSD2CLAUSE
os-sunnyvalley   1.2_1   652B   OPNsense   BSD2CLAUSE

any tip for me?

[ArminF]

hm..  as i had a backup i decided to reinstall zenarmor and elasticsearch.

no luck with that. DB does not start and system behaves very unstable.

[paschtin]

Same issue here. Cannot start the Elasticsearch database.
I did upgrade to 22.1 as well as the Zenarmor packages.

[ArminF]

twice reinstalled..

1st attempt with restore backup - elastic stopped
2ns with default config and just one interface - went OK
Started activating options, exclusion etc.
System went unstable and after reboot elasitc stopped.

besides is get this error as soon as i install sensei
[28-Jan-2022 10:38:50 Europe/Amsterdam] PHP Warning:  A non-numeric value encountered in /usr/local/sbin/carp_service_status on line 51

[hfvk]

Same issue here. OPNsense reinstalled Sensei after upgrading to 22.1 but Elasticsearch won’t start.

How do I force the reinstall?

[hfvk]

Altight, I reinstalled elasticsearch manually from CLI. Now it works, for some reason it was still on FreeBSD 12 ABI. Manual reinstall fixed and it is now 13 and everything is working fine

[ArminF]

Can you paste the CLI command code here?
So i can give it a try.

i installed three times through GUI.. failed all the time.

thanks Armin

[hfvk]

First, I did:
pkg deinstall elasticsearch5

After that:
pkg install elasticsearch5

You can use pkg info command to check the exact name of the package if the above does not work.

I am not sure if this was too "brute force" approach. I compared the installed packages with other identical systems I have running OK on 22.1 and the packages seem identical.

I upgraded multiple "identical" machines from 21.7 to 22.1 with same procedure. No idea what went differently here...

[ArminF]

Thanks hfvk

that did the trick.

pkg remove elasticsearh5
pkg install elasticsearch5

then enable and restore config&rules

[sy]

Hi all,

The repo is updated and it should be worked as expected.

[ArminF]

Thanks Sy,

but my issue came back right after reboot.
Elasticsearch does not start. System gehts unstable and reloads services every 1-2 minutes.

tried cli uninstall / install
removed elasticsearch user/group
removed left over files and directories.
reinstalled through GUI
still no luck.

Had 2 boxes in the office 1 was OK the other failed as well.