vfs.dev.read/write still broken on 1.8.x

[jeenam]

Longtime Zabbix user here getting aquainted with 1.8.2. In 1.6.x the vfs.dev.read/write feature was broken in the agent and I am wondering if this has been fixed in the latest 1.8.x tree.

From what I've been able to find on the forums you can monitor a single device (e.g. sda) but not a particular partition (e.g. sda4). The 1.4.x agent offered this feature. Why was it broken in later versions and not fixed?

What are the alternatives to this feature? Hard drive i/o monitoring is fairly essential for storage admins so the lack of this feature hurts.

tiv

[mister]

I hadn't look recently, but I changed my items when one of the agent versions stopped giving me values.

Well, I'm sure there could be a better way to do it, but I also liked a consolidated io value. I care about the overall i/o, most of the time, and there can be a *lot* of devices. I didn't see an easy way to do this with the built-in commands. Same approach would work for a single device.

In my case, I grabbed my own data using /proc/diskstats on linux, and kstat on solaris.

ex Solaris, all sd devices summed for read, returns bits/sec:
-> kstat for the data (nread for read data, nwritten for written data)
-> awk sums the values
-> Zabbix item: add a 8 multiplier to convert bytes to bits, select delta, and you get bits per sec.
-> similiar items set up for network, nfs, and san devices.

kstat -p -c disk 'sd:::nread' | awk '{FS="\t"; sum+=$2} END{OFMT="%d"; print sum}'

[jeenam]

Thank you for the reply. Will give it a go and report back.