BTRFS error on cache drive

[ipreferpie]

Overall:
    Device size:                   1.14TiB
    Device allocated:            645.09GiB
    Device unallocated:          524.89GiB
    Device missing:                  0.00B
    Used:                        641.25GiB
    Free (estimated):            525.27GiB      (min: 262.83GiB)
    Data ratio:                       1.00
    Metadata ratio:                   1.67
    Global reserve:              512.00MiB      (used: 0.00B)

Data,single: Size:640.00GiB, Used:639.62GiB
   /dev/mapper/sdh1      640.00GiB

Metadata,single: Size:1.00GiB, Used:317.94MiB
   /dev/mapper/sdh1        1.00GiB

Metadata,RAID1: Size:2.00GiB, Used:676.69MiB
   /dev/mapper/sdh1        2.00GiB
   /dev/mapper/sdi1        2.00GiB

System,single: Size:32.00MiB, Used:112.00KiB
   /dev/mapper/sdh1       32.00MiB

System,RAID1: Size:32.00MiB, Used:0.00B
   /dev/mapper/sdh1       32.00MiB
   /dev/mapper/sdi1       32.00MiB

Unallocated:
   /dev/mapper/sdh1      288.45GiB
   /dev/mapper/sdi1      236.44GiB

 

sdh is the 850EVO and sdi is the Crucial M4

[JorgeB]

If not done yet complete your backup and run the balance, but if you're still getting the " new device, device locked due to unknown error" you'll probably need to reformat the pool in the end and restore the data.

[ipreferpie]

ok thanks! I will complete the backup this time. Many thanks and will update soon. Just as a note, I was checking the SMART info on the SSDs. The 850 EVO has 0 reallocated and uncorrectable errors, but the Crucial M4 has 4096 reallocated and 5 uncorrectable. Funny thing is that it was the 850 that dropped earlier. Should I consider removing the Crucial M4 from the cache pool once everything is done? Looks like it may be dying

[JorgeB]

It might be a good idea, but if you keep both connect them on the onboard SATA as the LSI 2008 chipset doesn't support trim, also in my experience Samsungs are very picky with the SATA cable, make sure you're using a good quality cable.

[ipreferpie]

Totally had no idea that LSI doesn't do TRIM! Now that I know, I'll rewire my server since I have some SSDs on the IBM M1015. And thanks for the heads up on the Samsungs...only learning about that now too. All really valuable info to me

[ipreferpie]

looks like Cache 2 is in normal operation and unlocked after the balancing! Will run a scrub and then a parity check. If all goes well, I'll remove the Cache 1 Crucial M4 SSD from the pool due to 4096 unallocated errors. Glad things seem to be back on track and hope the verification goes well. Do you think these last steps sound good? Thank you so much for walking me through all this Here's the info:

 

Overall:
    Device size:           1.14TiB
    Device allocated:         643.06GiB
    Device unallocated:         526.92GiB
    Device missing:             0.00B
    Used:             640.65GiB
    Free (estimated):         527.15GiB    (min: 263.69GiB)
    Data ratio:                  1.00
    Metadata ratio:              2.00
    Global reserve:         512.00MiB    (used: 0.00B)

Data,single: Size:639.00GiB, Used:638.77GiB
   /dev/mapper/sdh1     639.00GiB

Metadata,RAID1: Size:2.00GiB, Used:961.73MiB
   /dev/mapper/sdh1       2.00GiB
   /dev/mapper/sdi1       2.00GiB

System,RAID1: Size:32.00MiB, Used:112.00KiB
   /dev/mapper/sdh1      32.00MiB
   /dev/mapper/sdi1      32.00MiB

Unallocated:
   /dev/mapper/sdh1     290.48GiB
   /dev/mapper/sdi1     236.44GiB

 

[JorgeB]

Pool looks normal now, steps are OK, just make sure backups are up date just in case.