zpool permanent errors

[vajonam]

  pool: MY_POOL

 state: ONLINE

status: One or more devices has experienced an error resulting in data

        corruption.  Applications may be affected.

action: Restore the file in question if possible.  Otherwise restore the

        entire pool from backup.

   see: http://www.sun.com/msg/ZFS-8000-8A

  scan: scrub in progress since Mon Jun  5 16:59:47 2017

    811G scanned out of 2.61T at 77.2M/s, 6h50m to go

    0 repaired, 30.38% done

config:

        NAME                       STATE     READ WRITE CKSUM

        MY_POOL                    ONLINE     16K     0     0

          mirror-0                 ONLINE     32K     0     0

            c0t50014EE262928278d0  ONLINE     32K     0     0

            c0t50014EE209912EF9d0  ONLINE     32K     0     0

errors: Permanent errors have been detected in the following files:

        mojave/dump:<0x1>

Not sure how to clear this error. I have done two scrubs and this hasn't helped. There is no filed called /mojave/dump at inode 0x1, where do I go from here.

Thanks.

[Andre Lue]

Not sure but since its a mirror I'd take a 50/50 guess at good device n boot w only 1 disk n try a read only import. If it still complains shutdown, swap mirror device n repeat steps w other.

How or what action caused this error?

--
You received this message because you are subscribed to the Google Groups "EON ZFS Storage" group.
To unsubscribe from this group and stop receiving emails from it, send an email to eonstorage+unsubscribe@googlegroups.com.
To post to this group, send email to eonst...@googlegroups.com.
Visit this group at https://groups.google.com/group/eonstorage.
For more options, visit https://groups.google.com/d/optout.

[vajonam]

No idea what caused the error. I did a scrub maybe a few days back when I re-did the case and added moved most of the drives over the mpt_sas 6gb/s controller. Then it did a scrub okay with no issues. I have monthly scrub that runs on each of my pools starting on the 1st of the month and this threw this up. 

nothing has changed, no - reboots etc. I am in the process of reading the data off into a temp pool and destroy and re-creating it as last resort. regardless I am going to backup the data. 

if I just have 1 disk in read only when will this error come up? 

This only complains on scrub not on import. 

On Monday, June 5, 2017 at 8:22:38 PM UTC-4, dre2kse wrote:

Not sure but since its a mirror I'd take a 50/50 guess at good device n boot w only 1 disk n try a read only import. If it still complains shutdown, swap mirror device n repeat steps w other.

How or what action caused this error?

On Jun 5, 2017 8:01 PM, "vajonam via EON ZFS Storage" <eonstorage+APn2wQeaGveeVtAFGo6StkzOM-yyat2Qk_yK0TpsH4zA9NaHLqHYt@googlegroups.com> wrote:

  pool: MY_POOL

 state: ONLINE

status: One or more devices has experienced an error resulting in data

        corruption.  Applications may be affected.

action: Restore the file in question if possible.  Otherwise restore the

        entire pool from backup.

   see: http://www.sun.com/msg/ZFS-8000-8A

  scan: scrub in progress since Mon Jun  5 16:59:47 2017

    811G scanned out of 2.61T at 77.2M/s, 6h50m to go

    0 repaired, 30.38% done

config:

        NAME                       STATE     READ WRITE CKSUM

        MY_POOL                    ONLINE     16K     0     0

          mirror-0                 ONLINE     32K     0     0

            c0t50014EE262928278d0  ONLINE     32K     0     0

            c0t50014EE209912EF9d0  ONLINE     32K     0     0

errors: Permanent errors have been detected in the following files:

        mojave/dump:<0x1>

Not sure how to clear this error. I have done two scrubs and this hasn't helped. There is no filed called /mojave/dump at inode 0x1, where do I go from here.

Thanks.

--
You received this message because you are subscribed to the Google Groups "EON ZFS Storage" group.

To unsubscribe from this group and stop receiving emails from it, send an email to eonstorage+...@googlegroups.com.

[Andre Lue]

Thats one way. You could try a guess at a error free half of the mirror also n then maybe try a replace w on the bad half.

To unsubscribe from this group and stop receiving emails from it, send an email to eonstorage+unsubscribe@googlegroups.com.

[vajonam]

Not sure how a read only import will help? can you explain that as I understand.

1. disconnect one of the drives add a new drive that will be error free

2. resilver everything 

3. run a scrub

try the other drive.

is that correct? 

On Monday, June 5, 2017 at 9:16:37 PM UTC-4, dre2kse wrote:

Thats one way. You could try a guess at a error free half of the mirror also n then maybe try a replace w on the bad half.

[Andre Lue]

1. disconnect one of the drives and boot
2, 3 good
If not ok put back previous removed drive n remove the one that just failed n repeat 2,3.

If none works, copy out to other pool n destroy.

If ok try adding back bad disk as a replace. You may have to create a temp zoop on it to blow the old metadata away before adding it back.

To unsubscribe from this group and stop receiving emails from it, send an email to eonstorage+unsubscribe@googlegroups.com.

[vajonam]

a,

Well, since one of the disks failed booting them worked regadless of which one.

But I did have mojave/dump zfs dataset that wasn't showing up in the fliesystem. do did a destroy on that dataset and scrub after which the errors have gone. not quite sure what caused that. but will keep an eye on the pool.

I do have ECC but no Xeon processor, I have ordered one that I will install to take advantage of the ECC to reduce silent corruption if any.

thanks for you pointers!