vfs.dev.read broken in Linux

[richlv]

Same problem here:
i have compiled agent 1.6.8, 1.8.4, 1.8.9 and 1.8.4 and when i execute the following command i get a [m|Collector is not started!] message

./zabbix_agentd -t vfs.dev.read[sda]

The same command on a fresh compiled zabbix agent 1.4.6 works perfect:

Does anybody know whats wrong? I monitor 250 hosts, and all the 1.6.x agents don't return any data.

you are doing it wrong (tm)
don't use -t/-p to test item keys, use telnet or zabbix_get against a running daemon

[jeenam]

They changed the syntax in versions above 1.4.x

Same problem here:
i have compiled agent 1.6.8, 1.8.4, 1.8.9 and 1.8.4 and when i execute the following command i get a [m|Collector is not started!] message

./zabbix_agentd -t vfs.dev.read[sda]

The same command on a fresh compiled zabbix agent 1.4.6 works perfect:

Does anybody know whats wrong? I monitor 250 hosts, and all the 1.6.x agents don't return any data.

Use this syntax for the item:

Key = vfs.dev.read[sda,sectors]
Type = Numeric (unsigned)
Data type = Decimal
Units = B/s
Use custom multiplier = 512
Store Value = Delta (speed per second)
Show Value = As is


Why they broke out of the box functionality is completely beyond me. It must be configured by hand, post install.

[Tristan]

i thought you gonna say that ;-)

ok:
zbx_agent version: 1.4.5
zabbix_get version: 1.8.4
command: ./zabbix_get -s myhostname -k vfs.dev.write[sda]
result: 9590584

zbx_agent version: 1.6.8
zabbix_get version: 1.8.4
command: ./zabbix_get -s myhostname -k vfs.dev.write[sda]
result: 10.266667

zbx_agent version: 1.8.4
zabbix_get version: 1.8.4
command: ./zabbix_get -s myhostname -k vfs.dev.write[sda]
result: 10.212

It looks like the format result for agent 1.4.5. differs from 1.6 and 1.8

[richlv]

most likely not the format, but default parameters used - sectors (which is a counter) vs <something> per second. try to pass parameter to specify type, value range should be more uniform then

[anjiytwok]

both read and write work just fine for me. does it work with zabbix_get for you ?
can you try on a 32bit box, just in case it's architecture specific ?

Hi,

I am also trying to monitor the Disk Read and write data on the host.

I have created the items like this.
1. For Writing data: vfs.dev.write[all,ops,avg1]
2. For Read data : vfs.dev.read[all,ops,avg1]

Problem is I am able see the write data, but read data is always zero(0)

What could be the problem, is there issue in the configuration. please see the item creation page and the data graph

please find the attached screens:

[tchjts1]

What version of the agent are you using?
I just tested this with 2.0.6 on my Zabbix DB server and am getting both values.

I changed my key from what you have though. I simply went with these:
vfs.dev.read[,ops,]
vfs.dev.write[,ops,]

I left out "all" on purpose, as that is the default.

(This is specific to agent 2.x)

https://www.zabbix.com/documentation/2.0/manual/config/items/itemtypes/zabbix_agent

As an FYI, also see the notes at the very bottom of that page.

[tchjts1]

Note that I have a mix of agent versions in my environment on the Linux boxes. Mostly 2.0.3 and 2.0.4, with 2.0.6 on my Zabbix boxes.

Looks like these counters may be broken on 2.0.3 agent (They throw "Not supported")
They do work on 2.0.4 and 2.0.6

[anjiytwok]

Hi tchjts1,

Thank for very quick response.

I am using the zabbix-2.0.4 vesrsion.

And I have tried using the Item for read: vfs.dev.read[,ops,]

Still there is no data. I could see only zero all the time.

But write data is clearly varying across the time.

I have no Idea on this why it is behaving.

And while I am going to through the zabbix forums, I found this link.

if i follow or modify the zabbix-agentd will it work, please see the below link:

http://www.muck.net/19/getting-hard-disk-performance-stats-from-zabbix

http://www.muck.net/19/getting-hard-disk-performance-stats-from-zabbix

Thank you once again.
:-)

[tchjts1]

Interesting. That info is about 5 years old, but I think it would still be fine as it is a userparameter. I'll give it a try when I get into the office later today and let you know how it works out.

I have a lot of these set up for Solaris (Disk IO stats) and would like to beef up my Linux templates as well.

[anjiytwok]

Hi tchjts1,

Did you get a chance to look at Linux templates, what could be the reason.

please reply me, when you find some time, Thank you.

[tchjts1]

So I was able to import the template by unchecking the "graphs" option. Import complained about that. Added the userparms to my agentd.conf and was able to pull data. Only issue I see is that you have to specify , [sdb], [sdc], [sda1], etc. There is no "all" functionality that I am aware of.

[anjiytwok]

Hi,

I have tried by adding the custom parameteres to zabbix-agentd config file

http://www.muck.net/19/getting-hard-...ts-from-zabbix

and restarted my zabbix agent

After that I have tried connecting throguh zabbix as well as on zabbix server by running the following command
1. On zabbix UI:
custom.vfs.dev.write.ops[sdb]
2. on zabbix server:
zabbix_get -s ec2-54-241-24-225.us-west-1.compute.amazonaws.com -p 10050 -k custom.vfs.dev.read.sectors[sdb]


I did not get any result for this. NOT supported excetpion

[tchjts1]

You may not have an sdb partition.

On your linux box you can run: sudo fdisk -l
Which will give you a list.

It'll look something like this

Disk /dev/sda: 73.2 GB, 73284976640 bytes
255 heads, 63 sectors/track, 8909 cylinders
Units = cylinders of 16065 * 512 = 8225280 bytes
   Device Boot      Start         End      Blocks   Id  System
/dev/sda1   *           1          13      104391   83  Linux
/dev/sda2              14        1580    12586927+  83  Linux
/dev/sda3            1581        8778    57817935   fd  Linux raid autodetect
/dev/sda4            8779        8909     1052257+   5  Extended
/dev/sda5           8779        8909     1052226   82  Linux swap / Solaris
Disk /dev/sdb: 73.2 GB, 73284976640 bytes
255 heads, 63 sectors/track, 8909 cylinders
Units = cylinders of 16065 * 512 = 8225280 bytes

[anjiytwok]

Hi tchjts1,

Sorry for late reply.

Here I have checked out the partition, using this command sudo fdisk -l


I did not get exactly what extract, why its not working for me.
Can you please guide me? how I need to fetch the read property data?

============


Disk /dev/xvda1: 107.4 GB, 107374182400 bytes
255 heads, 63 sectors/track, 13054 cylinders, total 209715200 sectors
Units = sectors of 1 * 512 = 512 bytes
Sector size (logical/physical): 512 bytes / 512 bytes
I/O size (minimum/optimal): 512 bytes / 512 bytes
Disk identifier: 0x00000000

Disk /dev/xvda1 doesn't contain a valid partition table

Disk /dev/xvdb: 450.9 GB, 450934865920 bytes
255 heads, 63 sectors/track, 54823 cylinders, total 880732160 sectors
Units = sectors of 1 * 512 = 512 bytes
Sector size (logical/physical): 512 bytes / 512 bytes
I/O size (minimum/optimal): 512 bytes / 512 bytes
Disk identifier: 0x00000000

Disk /dev/xvdb doesn't contain a valid partition table

[tchjts1]

The information is there.

So you have xvda1 and xvdb whereas I have sda1 and sda2.

Now if you go to that template that you imported and make the items with your information, attach that template to the server you got that info from, you should start seeing data coming in.

But this is where the issue lies, in that you have to be specific. There is no "all" capability that I am aware of for this. You would have to manage these items on a server-by-server basis. Very much a pain, and having a template is useless because managing items per-server defeats the purpose of a template.

My items look like this using sda1: (Change your "interval" to 60. the default of "10" is too often in my opinion)