Solved What are "video" disks?

PMc

PMc

Long time I wondered what all these rainbow colors might in which disks are sold nowadays. In former times we were used to some technical specifications or differenciate disks by, but now there are red, green, black and yellow disks.

Then there are video/surveillance disks, and these are only slightly more expensive than the desktop ones, but are rated for 24/7 operation, and some people say they might be suitable for NAS also.

So what are these? The answer is actually simple: disks without error correction.

After two of my SSD failed, I thought it might be better to run the pool with one leg on a spinning disk than with no redundancy at all. And this happens to be a video disk. It looks like this:

Code: 
  pool: im
 state: ONLINE
status: One or more devices has experienced an unrecoverable error.  An
        attempt was made to correct the error.  Applications are unaffected.
action: Determine if the device needs to be replaced, and clear the errors
        using 'zpool clear' or replace the device with 'zpool replace'.
   see: https://openzfs.github.io/openzfs-docs/msg/ZFS-8000-9P
  scan: resilvered 7.93G in 00:40:37 with 0 errors on Tue Jan  3 08:02:51 2023
config:

        NAME              STATE     READ WRITE CKSUM
        im                ONLINE       0     0     0
          raidz1-0        ONLINE       0     0     0
            ada5p1.elip1  ONLINE       0     0     0
            ada7.elip1    ONLINE       0     0     1
            ada1p1.elip1  ONLINE       0     0     0

Enterprise grade disk have a configuration where you can limit the error correction to a few seconds, in order to engage the raid error correction instead and not delay the raid. It looks like this in smartctl -x:

Code: 
SCT Error Recovery Control:
           Read:     80 (8.0 seconds)
          Write:     80 (8.0 seconds)

Desktop disks do not have this option, one cannot reduce the error correction. Video disks also do not have the option, but here one cannot activate the error correction. The result is then as shown above.
There are no errors in smart, there are no errors noticed by the controller, it just sends data with correct checksums, but bogus content.
Obviousely one can use such disks solely for videos where the software can repair/cover missing data and where a wrong pixel doesn't harm.
 
If a single device in raidz1 is returning bogus data, why is it an unrecoverable error? I doubt the problem is what you think it is.
 
bob2112 said:
If a single device in raidz1 is returning bogus data, why is it an unrecoverable error? I doubt the problem is what you think it is.
Click to expand...
AFAIK that is the normal message for chksum errors. It does state that applications are unaffected, so the ZFS has fixed the issue per raidz redundancy (which would not happen in non-redundant configuration).
But the individual device itself was apparently not able to recover (or, in this case, maybe didn't seriousely try to), otherwise the error would not have reached the ZFS layer.
The attached link explains this quite elaborately.
 
I confused the message with a different one. However, it's still only a single checksum error; it's not reasonable to infer anything about the design of that drive, or video disks in general, from it.
 
If that were the case, I wouldn't have.
It was 2 errors after resilver, one an hour later, and then another one every few hours.

Currently there are no more errors, after I changed a little bit:

Code: 
SCT Error Recovery Control:
           Read:    300 (30.0 seconds)
          Write:    300 (30.0 seconds)

It was like this before:

Code: 
SCT Error Recovery Control:
           Read: Disabled
          Write: Disabled
 
You must log in or register to reply here.
Back
Top