Zabbix Server - Debian 8.11 (
Zabbix Proxy - Debian 8.0 (4.9.35-v7+ #1014 SMP armv7l GNU/Linux)
Zabbix Database - Debian 8.11 | 10.2.17-MariaDB
vCenter Virtual Appliance - 6.5.0.22000

same for 3.4.14 - simillar situation was couple minor releases back - also with vmware items

tons of messages like this on the zabbix server

3.4.14 on CentOS affected as well.

most probably caused by: https://support.zabbix.com/browse/ZBX-14775

downgrading the proxy helped

@gofree: I don't have there proxy in the affected env., just the 3.4.14 server, and I'm not that eager to screw it up somehow more by downgrading it whole do 3.4.13.

in my environment proxies are doing all the metric collection - server is just for notifications ans metrics evaluation - so the logical step was to downgrade just the proxy which is monitoring vmware environemnt - i know its not recommended but it worked in 3.4.11 vmware issue as well

i dont want to downgrade the whole stack either

Can confirm: Debian 9 Stretch amd64 with Zabbix Proxy SQLite 3.4.14 affected.

Downgraded to 3.4.13 and dropped the SQLite DB in between. Fixed the problem.

Same Issue on Zabbix Appliance 4.0.0 Beta

MySQL DB

Edited. Solution in my next comment

To all with this issue, please see the following: https://kb.vmware.com/s/article/2107096

Please ensure the settings are set above in vCenter

Hi @lucanuk

My vCenter parameter is set to -1 (unlimited). I believe that downgrade is the solution (workaroud).

Confirmed

armv7l Proxy- after downgrade to 3.0.16 the problem solved.

amd64 Proxy - after downgrade to 3.0.21 also works.

Tks

Same issue here with Ubuntu 16.04 and Zabbix Server 3.4.14.

Corrected by adding the config.vpxd.stats.maxQueryMetrics key on my Vmware Vcenter 6.0 Appliance.

Thanks to @Nick Oram for the link above.

will this be handled in next release eg. can be ?

Same issue with Ubuntu 16.04 & ZS 3.4.14 & vCenter 6.5 (Windows). No solution by link of @Nick Oram

Adding config.vpxd.stats.maxQueryMetrics to my vCenter and setting it to the default (256 on my 6.5 installation) fixed the Zabbix 3.4.14 problem.  I assume downgrading Zabbix would have also "fixed" it.

Hi All
i solved.
https://kb.vmware.com/s/article/2107096
It was an effective solution for my environment.

*Ubuntu 16.04.5 LTS
*Zabbix Server 3.4.14
*postgreSQL 9.5
*vCenter5.5(vCSA)

MVekslers can you recall the previous values of maxQueryMetrics for vCenter and web.xml ?

Hi Michael

I confirmed it from a backup file.

Before:
   <context-param>
      <description>Specify the maximum query size (number of metrics) for a single report. Non-positive values disable the limit.</description>
      <param-name>maxQuerySize</param-name>
      <param-value>64</param-value>
   </context-param>

After:
   <context-param>
      <description>Specify the maximum query size (number of metrics) for a single report. Non-positive values disable the limit.</description>
      <param-name>maxQuerySize</param-name>
      <param-value>256</param-value>
   </context-param>

MVekslers Thanks for your reply. Did you change the value in vCenter (not in web.xml) ?

I will answer your question.

I have changed the two settings.

1.web.xml
2. Add vCenter parameter "config.vpxd.stats.maxQueryMetrics" key and the value.

that's not a solution, that's a work-around.

I've rolled back to 3.0.21. That worked. Hopefully this gets fixed soon in 3.0.23

Could be caused by ZBX-14775

Please understand that the

'config.vpxd.stats.maxQueryMetrics' is invalid or exceeds the maximum number of characters permitted**

error is what VMware returns when Zabbix tries to read config.vpxd.stats.maxQueryMetrics value to determine query limits. We are adding workaround to use default value 64 if this query fails, but better solution would be to configure VMware so correct limits could be obtained.

Available in:

• 3.0.23rc1 r85249
• 3.4.15rc1 r85250
• 4.0.0rc3 (trunk) r85251

Hi!

The official documentation has currently the following note:

To support datastore capacity metrics Zabbix requires VMware configuration vpxd.stats.maxQueryMetrics parameter to be at least 64. See also the VMware knowledge base article.

It is good. However, the mentioned article tells that the value "-1" for this parameter is a valid case:

7.To disable the limit, set a value to -1

The common sense suggests that "unlimited" is "more" than "64".
The current implementation of Zabbix server, however, disagree – it writes a big portion of messages like the following to the log:

 20523:20181122:161451.849 item "39373638-3935-5a43-4a38-30313436394e:vmware.hv.datastore.read[{$URL},{HOST.HOST},Tier1_Vesetas_VMFS5,latency]" became not supported: Cannot convert maxQueryMetrics from -1.
 20525:20181122:161451.889 item "39373638-3935-5a43-4a37-3431304b4a31:vmware.hv.datastore.read[{$URL},{HOST.HOST},Tier1_Vesetas_VMFS5,latency]" became not supported: Cannot convert maxQueryMetrics from -1.
 20526:20181122:161452.857 item "39373638-3935-5a43-4a38-30313436394e:vmware.hv.datastore.read[{$URL},{HOST.HOST},Tier1_DEAC_VMFS5,latency]" became not supported: Cannot convert maxQueryMetrics from -1.
 20526:20181122:161452.857 item "39373638-3935-5a43-4a37-3431304b4a31:vmware.hv.datastore.read[{$URL},{HOST.HOST},Tier1_DEAC_VMFS5,latency]" became not supported: Cannot convert maxQueryMetrics from -1.
 20525:20181122:161453.899 item "39373638-3935-5a43-4a38-30313436394e:vmware.hv.datastore.read[{$URL},{HOST.HOST},Tier1_Vesetas_vol2_VMFS5,latency]" became not supported: Cannot convert maxQueryMetrics from -1.

My opinion is that Zabbix server should take into account the possibility of such setting, and have a more stable behaviour.

Yes, it was fixed in ZBX-14858

wiper, thank you for a good news!
Probably, it should be useful to link both ZBX's as "related".

The changes of ZBX-14858 was merged with 3.0.26rc1.