Zabbix server service does not start "anymore"

[Eine Mail]

Hello...
I'm new here at Zabbix and slowly work my way through the functions...

Yesterday, in the middle of the business, the message suddenly came "Zabbix server is not running: The information displayed may not be up to date."



So just hopped on the Linux server and want to restart the service with a restart ... Unfortunately does not work!


Following data:
Linux system:
- Ubuntu 20.04 completely updated
- Firewall is deactivated

-> zabbix ran without problems for a good 2-3 weeks

Zabbix system: - Zabbix 5.0 LTS / Ubuntu 20.04 / PostgreSQL / Apache


- Check in /etc/zabbix/web/zabbix.conf.php whether everything is set correctly. (but it worked before too :-))
- with -> sudo service zabbix-server start / stop / restart tries to restart the service. (also restarted the server)
- with -> systemctl list-units --all --type = service --no-pager | grep zabbix -> the status checked

I am not good in Linux, the result is a bit confusing for me: -> Result: zabbix-server.service | loaded | activating | auto-restart | Zabbix server




To be honest, I have no idea where to look


I still found this error today, but unfortunately I cannot eliminate it or classify it






If someone has an idea, I would be very happy ...

Best regards Andy

[dimir]

The reason is most probably in the log file

/var/log/zabbix/zabbix_server.log

I assume you need to increase the value of some type of cache. Usually this looks something like this:

$ tail -n 100 /var/log/zabbix/zabbix_server.log
20244:20201022:135445.088 Starting Zabbix Server. Zabbix 5.0.4 (revision {ZABBIX_REVISION}).
20244:20201022:135445.088 ****** Enabled features ******
20244:20201022:135445.088 SNMP monitoring: YES
20244:20201022:135445.088 IPMI monitoring: NO
20244:20201022:135445.088 Web monitoring: YES
20244:20201022:135445.088 VMware monitoring: NO
20244:20201022:135445.088 SMTP authentication: YES
20244:20201022:135445.088 ODBC: NO
20244:20201022:135445.088 SSH support: NO
20244:20201022:135445.088 IPv6 support: YES
20244:20201022:135445.088 TLS support: YES
20244:20201022:135445.088 ******************************
20244:20201022:135445.088 using configuration file: /etc/zabbix/zabbix_server.conf
20244:20201022:135445.092 current database version (mandatory/optional): 05000000.0/05000002.0
20244:20201022:135445.092 required mandatory version: 05000000.0
20244:20201022:135445.098 server #0 started [main process]
20246:20201022:135445.099 server #1 started [configuration syncer #1]
20246:20201022:135445.107 __mem_malloc: skipped 0 asked 64 skip_min 18446744073709551615 skip_max 0
20246:20201022:135445.107 [file:dbconfig.c,line:96] __zbx_mem_malloc(): out of memory (requested 64 bytes)
20246:20201022:135445.107 [file:dbconfig.c,line:96] __zbx_mem_malloc(): please increase CacheSize configuration parameter
[...]

The message means you need to increase the value of CacheSize in /etc/zabbix/zabbix_server.conf .

[Eine Mail]

Thank you for your response.

I'm on a different track right now and will look at your proposal afterwards.

At the moment I see an entry "PidFile = / var / run / zabbix / zabbix_server.pid" in zabbix_server.conf,
but this file does not exist in the directory.

I'm currently looking for where this file could be.

I think that could be the problem ;-)

[dimir]

The file appears when Zabbix server starts successfully, since it fails to do so probably it shouldn't be there. I'd check the log file.

[Eine Mail]

Hey...

You're rigth.




I Change the CacheSize to 250M

Why was this parameter switched off by default??? :-)


A HUGE thank you :-)

[dimir]

We try to form the default configuration file so that user has all the information about each parameter. As to CacheSize:

### Option: CacheSize
# Size of configuration cache, in bytes.
# Shared memory size for storing host, item and trigger data.
#
# Mandatory: no
# Range: 128K-64G
# Default:
# CacheSize=8M

seems it's pretty straightforward. The default value (when commented out) is 8 MB, and you have the min and max supported values. We recommend to add new line after that section when you want to override:

### Option: CacheSize
# Size of configuration cache, in bytes.
# Shared memory size for storing host, item and trigger data.
#
# Mandatory: no
# Range: 128K-64G
# Default:
# CacheSize=8M
CacheSize=250M

[charliemav]

So this all makes perfect sense.... I added a bunch of Cisco devices on Friday with the templates and jumped up the demand on the server. No doubt the CacheSize was impacted. However, after reading this post, I adjusted the CacheSize parameters (many of the others as well), and still I can't get the Zabbix Server service to start. Log file still says out of memory.

sudo grep CacheSize -A 1 /etc/zabbix/zabbix_server.conf

### Option: VMwareCacheSize

# Size of VMware cache, in bytes.

--

VMwareCacheSize=32M




--

### Option: CacheSize

# Size of configuration cache, in bytes.

--

CacheSize=16G




--

### Option: HistoryCacheSize

# Size of history cache, in bytes.

--

HistoryCacheSize=256M




### Option: HistoryIndexCacheSize

# Size of history index cache, in bytes.

--

### Option: TrendCacheSize

# Size of trend write cache, in bytes.

--

TrendCacheSize=32M




### Option: TrendFunctionCacheSize

# Size of trend function cache, in bytes.

--

TrendFunctionCacheSize=32M




### Option: ValueCacheSize

# Size of history value cache, in bytes.

--

ValueCacheSize=250M


Went to the (seemingly) ridiculous number of 16G for the base CacheSize - still no dice. What am I missing here?