Pool Status Column Information

[Thousandbuckle]

Hi all, running TrueNAS-12.0-U7 and just replaced a HD and the OS is Resilvering right now as you can see in the upper left corner.

My question is what does the value of 9 for da7p2 under the Write column mean? This is the drive that was just replaced and I am concerned since it shows a value but none of the other drives do. I looked in the documentation but unable to find out what these columns mean.

 

[jgreco]

I would imagine it is from the output of "zpool status", write errors.

 

[Thousandbuckle]

So if there is "Any" value in the Read or Write or Checksum field then there is something wrong with the drive? I just ran the WD diagnostic software on the drive to make sure it was good prior to putting it in the array.

Is there any documentation in TrueNAS that states what these fields mean?

 

[jgreco]

Is there any documentation in TrueNAS that states what these fields mean?

I have no idea. It's sort of a "read the fine manual" thing if you want to. I don't typically read whatever basic documentation iXsystems produces since I don't need it, so, I don't know what documentation exists. Since the value appears to be inherited from ZFS, I'd look at the ZFS documentation anyways. It tends to be more technically rigorous than the iX documentation, which has usually been rewritten by a tech writer or two.

Oracle's documentation is available here:

https://docs.oracle.com/cd/E19253-01/819-5461/gbcve/index.html

 

[Ericloewe]

I was going to complain that pointing people at the Oracle docs could conceivably be dangerous and that we should refer them to the OpenZFS docs instead... but the man page for zpool status is not helpful and refers readers to the man page for zpoolconcepts, which also does not explain this output. Guess I need to learn how to write man pages for OpenZFS and send in a patch...

o if there is "Any" value in the Read or Write or Checksum field then there is something wrong with the drive?

Not strictly, it could be anywhere between CPU and disk. Typically, it's either the disk, controller, or the cables in between. SMART will narrow it down.

 

[jgreco]

but the man page for zpool status is not helpful and refers readers to the man page for zpoolconcepts, which also does not explain this output. Guess I need to learn how to write man pages for OpenZFS and send in a patch...

Somewhere around here I have a copy of Sun's Online: DiskSuite from the SunOS days (mid 1990's).

The documentation was a self-referential pile of claptrap that I think basically assumed you had a FE helping get you up to speed (or maybe were one of the developers?). Once you understood what was going on, you could then understand the thing, but there simply wasn't an equivalent to any of the ZFS intros. No Cyberjock slideshow or anything like that. Real frustrating stuff at times.

It's nice to know that the man pages for the complex storage systems we still rely on to this day are still an inscrutable mess.

 

[Ericloewe]

ZFS does a pretty good job, relatively speaking, but the standard is so low that it’s small comfort.