Ad Widget

**untergeek** · 10-12-2010, 18:07

Questions to help in troubleshooting:

What data type are these? e.g., Zabbix agent, Zabbix agent (Active), Zabbix trapper, simple, etc.

Have you tried doing a zabbix_get -s servername.example.com -k item.key against the server/agent to see if you can even get values (if they're agent based)

Perhaps that's a good first place to check, whether the values are even collectable manually.

**RaymondH** · 10-12-2010, 18:18

Thanks for the reply -

Since my last post it's a lot worse, so I'm not exactly sure what is going on.

I'll try the command you listed (if I can figure it out) and will let you know what happens. Kinda sad, it was running good for about 3 days after I fixed the last issue I had.

I don't know if it's because I'm using the appliance or it's just my overall lack of experience w/ Zabbix and Linux, but I'm just struggling to keep this running smoothly. It sure beats Nagios, just wish I could keep it running for more then a few days.

**untergeek** · 10-12-2010, 18:22

You'll do far better with an actual, compiled install. The appliance is great, but it's not for production (as they said).

**RaymondH** · 10-12-2010, 18:35

Yeah, that's my understanding. Except I'm to dumb to figure out how to compile the source even following directions. I get weird results, and then I can't figure out what to do after I "finish" (if I'm doing it right) the make install.

I've tried reading the directions in the Manual, some "how-to" guides for google, 2 different versions of a Linux install, and I just can't figure it out. So it was a case of: Use the appliance to play around w/ Zabbix, or just forget about it and look for something else.

The biggest challenge I'm having w/ this tool is that it's geared for people that know what they are doing in Linux/DB'es. It makes a lot of "we assume you know how to do X/Y/Z before you actually get to this step in the instructions". Which is fine, I understand, but it makes it very hard for a Windows guy to read between the lines and get it running/installed/configured.

Heck, took me 2 days to figure out I needed to convert the Appliance before I could use on my ESXi server.

Love this tool, don't get me wrong, it's awesome. Just seems to have a steep learning curve if you don't have Linux/DB experience.

Will get back to you on the other results - again, thanks for the reply/feedback.

**RaymondH** · 10-12-2010, 18:48

From my Zabbix server, it looks like I can use the zabbix-get command to pull the data.

hqzabbix:/usr/bin # ./zabbix-get -sx.x.x.x -p10050 -k"vfs.dev.read[sda,sectors]"
3364662

That being said - shouldn't it be working or am I missing something?

Thanks,

R

**untergeek** · 10-12-2010, 19:01

In theory. Performance implications are involved. If there are IO problems, you can have hang-ups. I just don't know enough about the appliance to know if you'd be suffering from this or not. It's complicated.

What is the interval between checks for these items? You could try decreasing the interval to be shorter.

Are these items Zabbix Agent or Zabbix Agent (Active)?

**RaymondH** · 10-12-2010, 19:16

Fair enough -

Let me do some more house keeping and figure out what I want my checks to be @; up to this point I've just left everything as it's default template value.

Will report back w/ my finding so that hopfully other that run into this can learn something.

Thanks again!

R

**RaymondH** · 10-12-2010, 23:02

:: Update ::

Whew....OK!

So for the better part of today, I went through one device at a time and disabled all UNSUPPORT Items/Triggers. (Is there an easy way to do this other than 1 at a time?)

This seems to have cleared up my issue; because about half way through the 89ish nodes I have monitored the Queue started to clear up and now I'm sitting at 0.

On the recommendation of Untergeek I also went through and adjusted my item timings.

Doing the above 2 things seemed to have done the trick....for now (dun dun dunnnnn).

Thanks for all the help, hopefully this post can help some other person who is experiencing the same issue!

**alixen** · 13-12-2010, 12:53

Hi,

Originally posted by RaymondH

So for the better part of today, I went through one device at a time and disabled all UNSUPPORT Items/Triggers. (Is there an easy way to do this other than 1 at a time?)

For Items:
In menu : Configuration -> Hosts -> Items
In Filter, select Status : Not supported
Clic on Filter.

At bottom of page drop down, select "Mass update" and clic on Go(nnn)
In mass update page:
Check Status checkbox
Set Status value to Disabled
Clic on Update

For Triggers, you can also use mass update but there is no filter, so you'll have to select all triggers by hand.

That's it

Regards,
Alixen

**RaymondH** · 13-12-2010, 14:34

Oh - you're awesome.

Thanks for the reply!

R

**qix** · 14-12-2010, 13:35

Hi Raymond,

I've also come across some large queue problems in the past.
In our case, it had to do with IO problems. The disks just couldn't keep up with the MySQL queries. Since you are on a VM Appliance, I suspect this is your problem as well.

Like untergeek replied earlier, setting a higher polling interval on your items will significantly decrease the IO load. (sorry for the late reply, doh.)

If you are new to Linux, just a few quick tips on seeing if you have an IO problem on your box:

Install atop, it is a top like application but it has a few bonuses. If your terminal supports it, it will colorcode lines that indicate a performance issue (like IO and Swapin/-out).
You could also use iotop, it is like top, but for io
If those are not available, iostat and/or vmstat might be able to help you identify if you are experiencing IO related trouble.

Just for being thorough, you need to get these tools in your VM appliance, not you ESXi server.

Just my 2cts worth, hope it helps,

Ad Widget

Queue Delays

Queue Delays

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment