Ad Widget

**Kirin2** · 30-08-2018, 14:42

How much have your increased cache size? Mb it's not enough? Also mb you need to check another paramaters related with cache size(I'm not sure, but you can try)

**dimir** · 30-08-2018, 16:41

Could you please attach the graph screenshot, which includes the time of crash (time bounds before and after the crash)?

**Mr.Fox** · 30-08-2018, 21:39

Thank you for your replies.

Originally posted by Kirin2

How much have your increased cache size? Mb it's not enough? Also mb you need to check another paramaters related with cache size(I'm not sure, but you can try)

I increased it from 4M to 32M then 128M and even 256M I think. What parameters are you thinking about ?

Originally posted by dimir

Could you please attach the graph screenshot, which includes the time of crash (time bounds before and after the crash)?

Also it's probably usefull to mention that i coudln't restart zabbix-server after the crash. I probably had to kill every zabbix server process but I restarted the machine to go faster.

Edit : decreased caches size yesterday and it seem to have taken longer to crash.

**Mr.Fox** · 04-09-2018, 01:41

Up.
My problem is still there, I also noticed the server was actually silently crashing sometimes but would automatically restart on it's own.

**dimir** · 04-09-2018, 14:11

I don't see vmware cache usage graph where it shows 2%.

**Mr.Fox** · 05-09-2018, 11:46

Originally posted by dimir

I don't see vmware cache usage graph where it shows 2%.

Last one, Zabbix VMware cache, % free 99,55%

**dimir** · 05-09-2018, 14:49

Ah, right, it's free. Is there enough memory on the server?

**Mr.Fox** · 05-09-2018, 17:08

About 3GB of available memory and 140 MB of free memory.

**Mr.Fox** · 10-09-2018, 09:45

Seems like the problem was coming from monitoring vCenter with the template. I noticed that the vCenter had significant higher CPU usage (+3Ghz) so I went back to monitoring via the ESXis and it's been steady for more than a week with lesser CPU usage on the ESXis.

Any idea on why that is ?

**LarryD** · 27-09-2018, 16:30

Originally posted by Mr.Fox

Seems like the problem was coming from monitoring vCenter with the template.

I'm having the same problem... can you elaborate on what you mean by this statement?

**Mr.Fox** · 28-09-2018, 13:33

Originally posted by LarryD

I'm having the same problem... can you elaborate on what you mean by this statement?

The problem i had was monitoring a vCenter with the built-in vmware template. When switching to monitoring via ESXi the problem went away.
I also found vCenter monitoring was using about 2-3Ghz of CPU usage on the vCenter.

Bumping up the vmcachesize way above what seemed to be needed utimately did the trick. What was wierd is that the cache size was 99.99% free but still it looked like sometime a very big chunk of data would come and made the poller crash (that wasn't happening with ESXi monitoring).
I increased the vm memory to 4GB and then increased the vmwarecachesize to 2G.
=> vCenter CPU usage dropped to normal level and cpu load of the Zabbix did the same.

Hope that helps

**vso** · 01-10-2018, 10:28

Hello! If you could please apply following patch, it could provide some insights.

Code:

Index: src/libs/zbxmemory/memalloc.c
===================================================================
--- src/libs/zbxmemory/memalloc.c    (revision 85057)
+++ src/libs/zbxmemory/memalloc.c    (working copy)
@@ -667,6 +667,9 @@
                 file, line, __function_name, (zbx_fs_size_t)size);
         zabbix_log(LOG_LEVEL_CRIT, "[file:%s,line:%d] %s(): please increase %s configuration parameter",
                 file, line, __function_name, info->mem_param);
+        zabbix_increase_log_level();
+        zabbix_increase_log_level();
+        zbx_mem_dump_stats(info);
         exit(EXIT_FAILURE);
     }

@@ -700,6 +703,9 @@
                 file, line, __function_name, (zbx_fs_size_t)size);
         zabbix_log(LOG_LEVEL_CRIT, "[file:%s,line:%d] %s(): please increase %s configuration parameter",
                 file, line, __function_name, info->mem_param);
+        zabbix_increase_log_level();
+        zabbix_increase_log_level();
+        zbx_mem_dump_stats(info);
         exit(EXIT_FAILURE);
     }

**vso** · 01-10-2018, 16:51

I am sorry but it looks like you have not applied patch or did not compile

**vso** · 01-10-2018, 19:37

What is your size of VMwareCacheSize configuration parameter?

Ad Widget

Please increase VMwareCacheSize configuration parameter

Please increase VMwareCacheSize configuration parameter

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment