ZFS power saving not possible due to spin up every 5 minutes

[runningman84]

I have installed talos on a new nas node (dxp2800) in my empty test cluster. I have created a raid1 pool instead of the two sata disks:

root@worker2:~# zpool list
NAME   SIZE  ALLOC   FREE  CKPOINT  EXPANDSZ   FRAG    CAP  DEDUP    HEALTH  ALTROOT
tank  10.9T  5.68G  10.9T        -         -     0%     0%  1.00x    ONLINE  -
root@worker2:~# zpool list
NAME   SIZE  ALLOC   FREE  CKPOINT  EXPANDSZ   FRAG    CAP  DEDUP    HEALTH  ALTROOT
tank  10.9T  5.68G  10.9T        -         -     0%     0%  1.00x    ONLINE  -
root@worker2:~# zpool status
  pool: tank
 state: ONLINE
config:

        NAME        STATE     READ WRITE CKSUM
        tank        ONLINE       0     0     0
          mirror-0  ONLINE       0     0     0
            sda     ONLINE       0     0     0
            sdb     ONLINE       0     0     0

root@worker2:~# zfs list
NAME          USED  AVAIL  REFER  MOUNTPOINT
tank         5.68G  10.8T   104K  /var/hddstorage
tank/public  5.68G  10.8T  5.68G  /var/hddstorage/public

The spin down setting for both sata drives were set using this command

root@worker2:~# hdparm -S 30 /dev/sda

/dev/sda:
 setting standby to 30 (2 minutes + 30 seconds)
root@worker2:~# hdparm -S 30 /dev/sdb

/dev/sdb:
 setting standby to 30 (2 minutes + 30 seconds)

With zpool iostat I can monitor the disk activity:

root@worker2:~# zpool iostat 1
              capacity     operations     bandwidth 
pool        alloc   free   read  write   read  write
----------  -----  -----  -----  -----  -----  -----
tank        5.68G  10.9T      0      1  4.30K  18.7K
tank        5.68G  10.9T      0      0      0      0

This is the full output recording 5 minutes:
https://gist.github.com/runningman84/e5d315b174202b2717594bbc9e34d9b6

The problem is that for some unkown reason every 5 minutes there is some io activty (7 or 8 write iops).

It is unclear to me what is the reason for these operations?

Is this caused by the ext-zpool-importer service? Is there some other talos service which writes to my disks?

[silpionmatze]

I also have this problem.

Exactly every 5 minutes

blktrace log:

  8,16   9        1     0.000000000  1470  A   W 469779104 + 24 <- (8,17) 469777056
  8,16   9        2     0.000000635  1470  Q   W 469779104 + 24 [z_wr_int_1]
  8,16   9        3     0.000004897  1470  G   W 469779104 + 24 [z_wr_int_1]
  8,16   9        4     0.000005568  1470  P   N [z_wr_int_1]
  8,16   9        5     0.000006317  1470  U   N [z_wr_int_1] 1
  8,16   9        6     0.000007213  1470  I   W 469779104 + 24 [z_wr_int_1]
  8,16   9        7     0.000012109  1470  D   W 469779104 + 24 [z_wr_int_1]
  8,16   0       14     0.000030481  7444  C   W 429716640 + 8 [0]
  8,16   0       15     0.000136773     0  C   W 469779104 + 24 [0]
  8,16  10        1     0.006915126  1568  A   R 44783864 + 8 <- (8,17) 44781816
  8,16  10        2     0.006915370  1568  Q   R 44783864 + 8 [txg_sync]
  8,16  10        3     0.006916296  1568  G   R 44783864 + 8 [txg_sync]
  8,16  10        4     0.006916754  1568  P   N [txg_sync]
  8,16  10        5     0.006916984  1568  U   N [txg_sync] 1
  8,16  10        6     0.006917191  1568  I   R 44783864 + 8 [txg_sync]
  8,16  10        7     0.006918855  1568  D   R 44783864 + 8 [txg_sync]

not using talos, so related to zfs zfs-2.2.4-1