USB Scrub Resulting in Errors. Normal?

[David3D]

This is the third USB boot device I've been through in as many months. Do they just not like the occasional scrub or am I just unlucky?

I'm not using bargin bin trash (or at least, I don't think so). I'm using Transcend drives and have never had an issue.

Daily run output:

Code:

Checking status of zfs pools:
NAME           SIZE  ALLOC   FREE  EXPANDSZ   FRAG    CAP  DEDUP  HEALTH  ALTROOT
PRIMARY       10.9T  4.07T  6.81T         -    19%    37%  1.00x  ONLINE  /mnt
SECONDARY     10.9T  3.72T  7.16T         -    21%    34%  1.00x  ONLINE  /mnt
freenas-boot  14.5G   505M  14.0G         -      -     3%  1.00x  ONLINE  -

  pool: freenas-boot
 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: http://illumos.org/msg/ZFS-8000-9P
  scan: scrub repaired 1K in 0h6m with 0 errors on Sat Sep 10 03:51:05 2016
config:

        NAME        STATE     READ WRITE CKSUM
        freenas-boot  ONLINE       0     0     0
          da12p2    ONLINE       0     0    18

errors: No known data errors

-- End of daily output --

Thanks for your input! :)

 

[Ericloewe]

or am I just unlucky?

No, most USB flash drives are just sucky.

Reinstall FreeNAS on a new drive and upload your config.

 

[Chris Tobey]

I've been having the same problem. They last one or two scrubs and then fatal error and I have to replace the USB drive.

Should I just switch to something else, or are their drives people recommend?

I'm using Patriot TAB drives currently.

 

[SweetAndLow]

Sandisk cruzer fit is the recommended USB around here. Mine is going 3 years strong.

Sent from my Nexus 5X using Tapatalk

 

[Spearfoot]

The SanDisk Cruzer Fit USB 2.0 are a good choice for FreeNAS boot drives. I used them for over a year w/o problems before switching to SSDs.

For some reason, USB 2.0 devices seem to hold up better than 3.0.

 

[Chris Tobey]

Done, purchased two off Amazon, should be here tomorrow :)

 

[David3D]

Interesting re the 2.0 vs. 3.0 drives. I'll follow on and buy a couple of the SanDisk Cruizers. For some reason, they always came off a bit tacky/cheap to me. I think it's the silly name they gave it. Seems like I was wrong.

 

[Chris Tobey]

The SanDisk Cruzer Fit 2.0 drives have been working well for the last 4+ months. Much better than the other drives.

 

[DrKK]

Sandisk Cruzer Fit, on 24/7 in my FreeNAS, 3+ years now.

Never one bit out of place.

 

[Redcoat]

Sandisk Cruzer Fit, on 24/7 in my FreeNAS, 3+ years now.

Never one bit out of place.

I bought a couple of these for my Dell C2100 FS12-TY new acquisition based on all the strong recommendations here. Lost one at 1st scrub. Replaced in kind, so far so good after 2 scrubs.

 

[Redcoat]

I bought a couple of these for my Dell C2100 FS12-TY new acquisition based on all the strong recommendations here. Lost one at 1st scrub. Replaced in kind, so far so good after 2 scrubs.

Just lost another at about 4 weeks lifetime.

 

[SweetAndLow]

Just lost another at about 4 weeks lifetime.

USB devices die all the time. Make sure you not have your system dataset and logs writing to the USB. My USB has see zero write since the last upgrade I did so nothing can really hurt it. If you're still having issues switch to a small ssd, they cost $50.

Sent from my Nexus 5X using Tapatalk

 

[Redcoat]

USB devices die all the time. Make sure you not have your system dataset and logs writing to the USB. My USB has see zero write since the last upgrade I did so nothing can really hurt it. If you're still having issues switch to a small ssd, they cost $50.

Sent from my Nexus 5X using Tapatalk

Thanks for the suggestion. I assume I would know if those writes were taking place by apparent size of boot load and its increase. I'm sure they are not, but I'm going to watch it.
I've got several well-run-in SATA laptop hard drives (replaced by SSD's) on the desk here in front of me and open cables in the Dell from motherboard drive connectors - is there any good reason not to consider using them instead of the USB's? I'm familiar with the use of SSD's for FreeNAS boot - my Mini came with one - but the laptop drives are "free"...

 

[SweetAndLow]

Here is an example of what my usb read/write usage looks like over the last week. Basically a read of 200 bytes daily and the writes are when I made a config change recently. Other than that zero activity.

 

[Redcoat]

Here is an example of what my usb read/write usage looks like over the last week. Basically a read of 200 bytes daily and the writes are when I made a config change recently. Other than that zero activity.
View attachment 16379

Well, I think you nailed it - I appear to have have writes every hour:

Systemdataset is on the pool, not the USB, so can you advise me how to find what is writing there, please?

 

[SweetAndLow]

How about your graphing data. Called rrd or something? That should be on your pool also.

Sent from my Nexus 5X using Tapatalk

 

[Redcoat]

How about your graphing data. Called rrd or something? That should be on your pool also.

Sent from my Nexus 5X using Tapatalk

Yes, they're in var/db/collectd/rrd/localhost.

So, now I ask myself what happens every hour on both my FreeNAS systems - and the answer is "to my current knowledge, nothing".

On my two-user network here the only scheduled task I know about is an hourly transfer of changed data from my primary Win10 desktop to an SMB share on FreeNASMini. Nothing currently writes to DellFreeNAS except a snapshot replication task from FreeNASMini. Snapshots on that run every 12 hours every day.

Both machines have the hourly boot USB write phenomenon. The Dell has two Sandisk Cruzer Fit 32GB USB's in mirror. The FreeNASMini has a 15 GB SATADOM.

So what's writing to my boot drives and how do I find it? Why do I have frequent writes when others do not - what's different with my configuration - what have I, could I have, set that caused this behavior (system dataset on both machines is in the pool, and the RRD data is not being saved)? That's presently above my pay grade, but I hope to get there with help from here - thanks in advance for any data, instructions, pointers, that anyone is willing to share.

While I do want to understand what's going on (and change the behavior if possible/necessary) , I am working on getting a SATADOM for the Dell as boot drive (probably Innodisk). Do I get two and mirror or stick with one - I didn't identify a preferred course in the postings I read last night?

 

[SweetAndLow]

What is the output of mount? Did you actually check the box to store rrd data on the pool?

Sent from my Nexus 5X using Tapatalk

 

[Redcoat]

What is the output of mount? Did you actually check the box to store rrd data on the pool?

Sent from my Nexus 5X using Tapatalk

No, I did not check the box to put RRD data on the pool - I have now done so - I infer this will either alter a default or provide a check if the write behavior stops?

Here's what mount returns:
[root@Dell] ~# mount
freenas-boot/ROOT/9.10.2-U2 on / (zfs, local, noatime, nfsv4acls)
devfs on /dev (devfs, local, multilabel)
tmpfs on /etc (tmpfs, local)
tmpfs on /mnt (tmpfs, local)
tmpfs on /var (tmpfs, local)
freenas-boot/grub on /boot/grub (zfs, local, noatime, nfsv4acls)
fdescfs on /dev/fd (fdescfs)
tank/Minibackup/jails/.warden-template-standard on /mnt/Volume1/jails/.warden-template-standard (zfs, local, nfsv4acls)
tank on /mnt/tank (zfs, local, nfsv4acls)
tank/Minibackup on /mnt/tank/Minibackup (zfs, local, nfsv4acls)
tank/Minibackup/Volume1 on /mnt/tank/Minibackup/Volume1 (zfs, local, read-only, nfsv4acls)
tank/Minibackup/jails on /mnt/tank/Minibackup/jails (zfs, local, nfsv4acls)
tank/Minibackup/jails/plex on /mnt/tank/Minibackup/jails/plex (zfs, local, nfsv4acls)
tank/Minibackup/media on /mnt/tank/Minibackup/media (zfs, local, nfsv4acls)
tank/Minibackup/private on /mnt/tank/Minibackup/private (zfs, local, nfsv4acls)
tank/Minibackup/windowsshare on /mnt/tank/Minibackup/windowsshare (zfs, local, nfsv4acls)
tank/Minibackup/windowsshare/Internas1 on /mnt/tank/Minibackup/windowsshare/Internas1 (zfs, local, nfsv4acls)
tank/Minibackup/windowsshare/Scans on /mnt/tank/Minibackup/windowsshare/Scans (zfs, local, nfsv4acls)
tank/jails on /mnt/tank/jails (zfs, local, nfsv4acls)
tank/.system on /var/db/system (zfs, local, nfsv4acls)
tank/.system/cores on /var/db/system/cores (zfs, local, nfsv4acls)
tank/.system/samba4 on /var/db/system/samba4 (zfs, local, nfsv4acls)
tank/.system/syslog-ea02119b0df4495ba64ec1dbdd61ed06 on /var/db/system/syslog-ea02119b0df4495ba64ec1dbdd61ed06 (zfs, local, nfsv4acls)
tank/.system/rrd-ea02119b0df4495ba64ec1dbdd61ed06 on /var/db/system/rrd-ea02119b0df4495ba64ec1dbdd61ed06 (zfs, local, nfsv4acls)
tank/.system/configs-ea02119b0df4495ba64ec1dbdd61ed06 on /var/db/system/configs-ea02119b0df4495ba64ec1dbdd61ed06 (zfs, local, nfsv4acls)
[root@Dell] ~#

Thank you for your interest and engagement! I do appreciate it and the insights I am gaining!

 

[SweetAndLow]

No, I did not check the box to put RRD data on the pool - I have now done so - I infer this will either alter a default or provide a check if the write behavior stops?

Here's what mount returns:
[root@Dell] ~# mount
freenas-boot/ROOT/9.10.2-U2 on / (zfs, local, noatime, nfsv4acls)
devfs on /dev (devfs, local, multilabel)
tmpfs on /etc (tmpfs, local)
tmpfs on /mnt (tmpfs, local)
tmpfs on /var (tmpfs, local)
freenas-boot/grub on /boot/grub (zfs, local, noatime, nfsv4acls)
fdescfs on /dev/fd (fdescfs)
tank/Minibackup/jails/.warden-template-standard on /mnt/Volume1/jails/.warden-template-standard (zfs, local, nfsv4acls)
tank on /mnt/tank (zfs, local, nfsv4acls)
tank/Minibackup on /mnt/tank/Minibackup (zfs, local, nfsv4acls)
tank/Minibackup/Volume1 on /mnt/tank/Minibackup/Volume1 (zfs, local, read-only, nfsv4acls)
tank/Minibackup/jails on /mnt/tank/Minibackup/jails (zfs, local, nfsv4acls)
tank/Minibackup/jails/plex on /mnt/tank/Minibackup/jails/plex (zfs, local, nfsv4acls)
tank/Minibackup/media on /mnt/tank/Minibackup/media (zfs, local, nfsv4acls)
tank/Minibackup/private on /mnt/tank/Minibackup/private (zfs, local, nfsv4acls)
tank/Minibackup/windowsshare on /mnt/tank/Minibackup/windowsshare (zfs, local, nfsv4acls)
tank/Minibackup/windowsshare/Internas1 on /mnt/tank/Minibackup/windowsshare/Internas1 (zfs, local, nfsv4acls)
tank/Minibackup/windowsshare/Scans on /mnt/tank/Minibackup/windowsshare/Scans (zfs, local, nfsv4acls)
tank/jails on /mnt/tank/jails (zfs, local, nfsv4acls)
tank/.system on /var/db/system (zfs, local, nfsv4acls)
tank/.system/cores on /var/db/system/cores (zfs, local, nfsv4acls)
tank/.system/samba4 on /var/db/system/samba4 (zfs, local, nfsv4acls)
tank/.system/syslog-ea02119b0df4495ba64ec1dbdd61ed06 on /var/db/system/syslog-ea02119b0df4495ba64ec1dbdd61ed06 (zfs, local, nfsv4acls)
tank/.system/rrd-ea02119b0df4495ba64ec1dbdd61ed06 on /var/db/system/rrd-ea02119b0df4495ba64ec1dbdd61ed06 (zfs, local, nfsv4acls)
tank/.system/configs-ea02119b0df4495ba64ec1dbdd61ed06 on /var/db/system/configs-ea02119b0df4495ba64ec1dbdd61ed06 (zfs, local, nfsv4acls)
[root@Dell] ~#

Thank you for your interest and engagement! I do appreciate it and the insights I am gaining!

I think the graphs write data no matter what. My theory is they write to the UBS by default then erase them at a certain point. Now I would watch your USB writes and see if the behavior changes.

Sent from my Nexus 5X using Tapatalk