|
Hello!
Please attach full logs from Zabbix server pod (Logs tab).
|
|
zabbix_server_POD.log
|
|
zabbix-operator-certified-5898d64d5d-8szj7-ansible.log
|
|
I try another strategy (Zabbix Appliance) and works. I understand the problem is with Zabbix full strategy.
|
|
It looks like you used claim volume with already existing DB, probably from previous tests:
** Creating 'zabbix' user in MySQL database
** Database 'zabbix' already exists. Please be careful with database COLLATE!
** Table 'zabbix.dbversion' already exists.
** Preparing Zabbix server configuration file
Could you try completely new volume claim and check log file again?
|
|
I built a new environment from scratch and the same issue persists.
Again, If I build an Appliance strategy, everything works as expected.
Do you have a specific guide/reproducer for this Zabbix full strategy ? I would be more than happy to test and provide a feedback to you.
|
|
It must be simple, no any additional prerequisites! Terrible sorry for such issue! We will check it and solve as soon as possible!
|
|
I'm just adding some information to show this is working on the Appliance strategy.
zabbix-appliance-appliance-59dc6c5858-lvxkn-zabbixappliance.log 
|
|
According log file appliance looks correct. Do you have any issues with Appliance as well?
|
|
How I said before, Appliance strategy is working as expected. My problem is just with Zabbix Full strategy.
|
|
Just to set up my expectation: Have you ever a chance to take a look on that? Do you have an ETA ?
|
|
Hello Alessandro, I hope this or next week maximum!
It would be great if you can attach MySQL database log file from Zabbix Full strategy after scratch install to empty volume claim.
|
|
Hello Alessandro!
I'm terrible sorry for a delay with response. I can not reproduce the problem. I followed the next procedure:
1. Created secrets:
kind: Secret
apiVersion: v1
metadata:
name: zabbix-full-secrets
data:
mysql_root_password: emFiYml4X3N1cGVyX3Jvb3Q=
mysql_zabbix_password: emFiYml4X3VzZXJfcGFzc3dvcmQ=
mysql_zabbix_username: emFiYml4
type: Opaque
2. Volume claim:
% kubectl get pvc
NAME STATUS VOLUME CAPACITY ACCESS MODES STORAGECLASS AGE
mysql-zabbix Bound pv0007 100Gi RWO,ROX,RWX 9m
% kubectl get pv | grep pv0007
pv0007 100Gi RWO,ROX,RWX Recycle Bound zabbix/mysql-zabbix 13d
3. Zabbix Full:
spec:
java_gateway_size: 1
server_env:
general:
zabbix_mysql_volumeclaim: mysql-zabbix
zabbix_mysqlsecret: zabbix-full-secrets
java_gateway:
debug_level: info
start_pollers: 5
timeout: 3
server:
start_discoverers: 1
trend_cache_size: 4M
start_trappers: 5
timeout: 4
log_slow_queries: 0
cache_update_frequency: 60
start_java_pollers: 5
start_alerters: 3
start_db_syncers: 4
start_preprocessors: 3
start_ipmi_pollers: 0
history_cache_size: 16M
start_pollers: 5
start_timers: 1
unreachable_period: 45
value_cache_size: 8M
start_http_pollers: 1
proxy_config_frequency: 3600
vmware_frequency: 60
vmware_perf_frequency: 60
trapper_timeout: 300
max_housekeeper_delete: 5000
unavailable_delay: 60
history_index_cache_size: 4M
vmware_timeout: 10
vmaware_cache_size: 8M
start_pingers: 1
unreachable_delay: 15
proxy_data_frequency: 1
debug_level: 3
start_proxy_pollers: 0
start_escalators: 1
housekeeping_frequency: 1
cache_size: 8M
start_pollers_unreachable: 1
web:
max_execution_time: 300
max_input_time: 300
memory_limit: 128M
post_max_size: 16M
server_name: Kubernetes installation
session_name: zbx_sessionid
timezone: Europe/Riga
upload_max_filesize: 2M
version: latest
web_size: 2
Finally everything is working fine:
Zabbix server pod:
|
|
Still not working, but let's move one step back. My environment is AWS + Openshift 4.4 + Openshift Container Storage (4.4) for storage backend (RBD).
Again, I'm doing the same steps to Zabbix Appliance and is working as expected. Maybe, OCS is not supported as storage backend?
I see the installer is not creating Zabbix Database:
My Pods:
NAME READY STATUS RESTARTS AGE
zabbix-full-db-85c5f6fb69-s24pt 1/1 Running 0 84s
zabbix-full-java-gw-0 1/1 Running 0 77s
zabbix-full-server-5d89b585-v9qkf 0/1 CrashLoopBackOff 2 81s
zabbix-full-web-0 1/1 Running 0 74s
zabbix-full-web-1 1/1 Running 0 47s
zabbix-operator-certified-6549c99b5f-46f25 2/2 Running 0 24m}}
|
|
According the output the DB schema and tables already exist:
** Creating 'zabbix' user in MySQL database
** Database 'zabbix' already exists. Please be careful with database COLLATE!
** Table 'zabbix.dbversion' already exists.
If OCS is transparent for MySQL server, I do not see any problems with the storage.
Is it possible to pass DEBUG_MODE="true" env variable to Zabbix server container and then attach extended log?
|
|
My hundred installation from scratch is showing the same result.
The reproducer in debug mode is available here: https://paste.centos.org/view/52f83a1d
Sorry, but I'd like to know:
1. What is the storage used by you ?
2. Are you using Vanilla Kubernetes or Openshift in your reproducer ?
|
|
I tried to use Openshift of course 
I'm thinkg in you case some issue with InnoDB storage engine. So could you pass DEBUG_MODE="true" env variable to Zabbix server pod and show us logs? Also I see Zabbix frontend is running, are you able to use it?
|
|
We have noticed that this issue has not been updated for a long time. May we assume that this case is resolved? If it is not and requires additional work, please comment on the ticket and we will respond accordingly. Alternatively, we're going to change the state of the issue to Resolved in 14 days if we don't hear back from you.
|
|
Issue was not solved. Unfortunately, I'm not able to reproduce the issue again cause I don't have the same env anymore. I will do it soon.
|
|
zabbix_DB_queries.log
|
|
zabbix-full-server-7889f764d6-ljwzp-zabbix-server.log
|
|
zabbix-full-db-59dfddbc7-89bgs-zabbixdb.log
|
|
I added zabbix server POD log on debug mode and more logs about DB that I think you should consider.
The route for the service is configured and after try to access ZBX frontend I'm facing an issue. 
My current status still the same:
# oc get pods --namespace zabbix
NAME READY STATUS RESTARTS AGE
zabbix-full-db-59dfddbc7-89bgs 1/1 Running 0 3h
zabbix-full-java-gw-0 1/1 Running 0 3h
zabbix-full-server-7889f764d6-ljwzp 0/1 CrashLoopBackOff 32 145m
zabbix-full-server-c5cb6c65f-chmfp 0/1 CrashLoopBackOff 33 148m
zabbix-full-web-0 1/1 Running 0 3h
zabbix-full-web-1 1/1 Running 0 179m
zabbix-operator-certified-5f8d8d7f79-2b78h 2/2 Running 0 17h
|
|
Hello!
We found the issue, the operator will be updated during this week. Currently I can recommend you provide more resources to Zabbix Full.
Kind regards,
Alexey.
|
|
That's definitely a good news! What do you mean about more resources? Please, could you elaborate more ?
I know you are moving to update the operator this week, but I would like to still working on it to understand.
Thanks!
|
|
We have fixed the issue and upgraded operator to 5.0. Please try new version!
|
|
I have upgraded the Operator as your sugestion, re-deployed Zabbix Full and worked as expected, but just to confirm that we don't have a bug anymore, I did:
- Delete zabbix-full cluster, volumes and secret
- Delete Zabbix Operator
- Installed Zabbix Operator LTS from scratch
- Create a new secret and volume claim
- Deploy Zabbix Full again
I tried this many times and unfortunately is not working. The issue is the same of the beginning.
Do you have any specific suggestion about parameters to deploy Zabbix full strategy to avoid these issues?
You have mention something about resources but this is not clear for me.
Thanks!
|
|
Now, I tested again and works. I really do not know why because the procedure is the same. Maybe this could be related to the resources or my cluster.
POD events is showing:
Startup probe failed: ERROR 2002 (HY000): Can't connect to local MySQL server through socket '/var/lib/mysql/mysql.sock' (2)
So, I understand this could be an intermittent issue.
|
|
Yes, it is related to node resources + previously was incorrect specification to use only livenessProbe. Currently added as well:
startupProbe:
tcpSocket:
port: 10051
initialDelaySeconds: 15
timeoutSeconds: 3
periodSeconds: 5
successThreshold: 1
failureThreshold: 40
Which can help with initial DB schema creation.
Can we consider the issue resolved?
PS: If you have 5.0.3 version, it is completely new operator with included bug fix
|
Generated at Mon Apr 28 11:08:41 EEST 2025 using Jira 9.12.4#9120004-sha1:625303b708afdb767e17cb2838290c41888e9ff0.
Captura de tela de 2020-06-12 19-40-25.png
zabbix-appliance-appliance-59dc6c5858-lvxkn-zabbixappliance.log
