ElasticSearch 7.5.0 still having troubles?

Hey folks,

is anyone still having troubles after upgrading to ES 7.5.x? I’ve seen numerous topics on here with people having different problems.

Because i just did an update and everything seems to be working pretty fine. I’m just afraid i’m overlooking something :smiley:.

cheers

1 Like

ES 7.5 works for me aswell

1 Like

I’ve got a general Question about the Elasticsearch Update. Are there any know problems with the Update? How i can update Elasticsearch? Is a

apt install elasticsearch

enough if i add the newest Elasticsearch Repository?

That’s what i’m trying to find out in this thread. I did an update and didn’t have any troubles yet. @x-f didn’t have any troublers either.

Depending on your current version you have to follow some path. Just googling “upgrade elasticsearch” returns this site as first result: https://www.elastic.co/guide/en/elasticsearch/reference/current/setup-upgrade.html

Stick to this and you’re fine. In addition to that remember that ES 7.5.0 doesn’t like indices build with an earlier version. You’d need to reindex them or the simple way: just delete all indices before upgrading and let zammad rebuild them after the upgrade. This can take a lot of time depending on your env (takes about 2h on my system) and while the indices are deleted a few functions won’t be available in zammad (search, users in the backend, autofill for mail contacts,…).

cheers

@svnr-dvnkln
Thank you for this Informations. I must update from 5.6 to the newest Version. I hope i found some time in the Holiday Time to update Elasticsearch.

Actually i got some Indices Problem (some Users won’t be found), so the Update could solve this.

Can i ask how big your env is (How much Agents are working at the same time with Zammad)?

Concurrent agents aren’t relevant for your indices. What matters is the index store.size and the docs.count.

You can check those by firing this command on your es server:
curl -X GET "localhost:9200/_cat/indices/zammad*?v&s=index&pretty"

Or on any remote server that is allowed to connect to your es server but you’d have to adjust localhost (or even the port depending on your conf) in that case. My indices look like this currently (our env is rather small):

index                                                 docs.count store.size
zammad_production                                             30    508.9kb
zammad_production_chat_session                                 0      1.2kb
zammad_production_cti_log                                      0      1.2kb
zammad_production_knowledge_base_answer_translation            3     22.5kb
zammad_production_knowledge_base_category_translation          2       10kb
zammad_production_knowledge_base_translation                   1      6.5kb
zammad_production_organization                                67      1.5mb
zammad_production_stats_store                                796        2mb
zammad_production_ticket                                   55939    292.6mb
zammad_production_user                                      1041      2.9mb

Hint: As you probably can tell (at least for the big players here ticket/user) the docs.count is equal to the actual number of tickets/users in zammad. If you have no indeces build yet you can just ask zammad directly via:
zammad run rails r 'puts Ticket.all.count'
zammad run rails r 'puts User.all.count'

cheers

@svnr-dvnkln

You Indices are looking completly different from mine. If i put you Command on the Server iam getting this:

health status index uuid pri rep docs.count docs.d eleted store.size pri.store.size
yellow open zammad_production 415qksLgQLGdF9h43hWbIg 5 1 11734 168 133.2mb 133.2mb

How i can see the Status of the Index is yellow, so it looks like that something is not okay. Why you got so many Indices and i got only one?

Oh okay sorry can’t help you with this.
Zammad usually creates these indices via: zammad run rake searchindex:rebuild (at least on my setup). It’s nothing i do manually.

Maybe someone from the zammad team can explain why they’re different on our env.
What’s your setup? Mine is:

  • Used Zammad version: 3.2.0-1575624531.d79c85cf.stretch
  • Used Zammad installation source: package
  • Operating system: deb 9

Mine are yellow too btw. That’s perfectly fine for me because it just means i could lose my data if something goes downtown with my node since i’m not running a cluster. That’s a risk i’m willing to take because in that case i just need to let zammad rebuild my indices. If you want it to be displayed as green you could set the number of replicas to 0 manually. That way es shouldn’t complain anymore.

P.S.: Maybe it looks different on your side because we have different es versions, just a guess though because i didn’t have a look at it with es 5.6.

cheers

@svnr-dvnkln

Iam using the news Zammad 3.2 Version on Ubuntu 18.04LTS, Installation Source is Package.

Maybe @MrGeneration can give us a answer why we got a different number of indices.

No I can’t :slight_smile:
Normally the first part before _tickets etc is different, but the last part is always the same (you should have indicies like Devin has.

Regarding the other issues:
Zammad 3.2 supports Elasticsearch 7.5 in it’s current stable, Zammad earlier 3.2 does not because of technical reasons.

This topic was automatically closed 41 days after the last reply. New replies are no longer allowed.