[SOLVED] Daily backups - spinning disks consideration

Kardell · 2024-05-30 23:58:11

Hi All,
I hope you are well.
I think about building a disk array for some backups using spinning disks or maybe even SSDs.
Can you control the storage hardware so it does not consume electricity, spinning while idle?
This would be a storage that is not needed all the time perhaps mounted and dismounted everyday just for the backups.
Electricity consumption is one thing but also wear if it's to do with mechanical storage device.
Let me know.
Best regards

Last edited by Kardell (2024-05-31 00:14:35)

gromit · 2024-05-31 00:04:29

Yes this is possible, see https://wiki.archlinux.org/title/Hdparm … figuration for reference

My NAS right now:

$ hdparm -C /dev/sd{a,b}

/dev/sda:
 drive state is:  standby

/dev/sdb:
 drive state is:  standby

# initially set with the following
$ hdparm -S 120 /dev/sd{a,b}

Kardell · 2024-05-31 00:14:16

That's awesome, thank you!

cryptearth · 2024-05-31 05:45:30

You may also check if the drives and the controller support power up in stand by. This keeps the disk from spinning up on power on but only when they're requested by the controller.
This can also help to bring up a big array with a smaller power supply which can keep up with the constant load but doesn't have enough oomph for the high inrush current when all drives try to start all at once.
Note: It depends on both the drives and the controller - if the controller doesn't support that feature you xan end up with a "dead" drive unless you connect it to a system which does support it to bring it back to live.

Kardell · 2024-08-11 03:31:04

cryptearth wrote:

You may also check if the drives and the controller support power up in stand by. This keeps the disk from spinning up on power on but only when they're requested by the controller.
This can also help to bring up a big array with a smaller power supply which can keep up with the constant load but doesn't have enough oomph for the high inrush current when all drives try to start all at once.
Note: It depends on both the drives and the controller - if the controller doesn't support that feature you xan end up with a "dead" drive unless you connect it to a system which does support it to bring it back to live.

Just an update re this. I have two of these 3.5" NAS hard drives: Seagate IronWolf Pro but I cannot see them (the driver) supporting the standby or spindown.
hdparm gives me:

hdparm -C /dev/sd{b,c}

/dev/sdb:
SG_IO: bad/missing sense data, sb[]:  f0 00 01 00 50 40 81 0a 00 82 00 00 00 1d 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
 drive state is:  unknown

/dev/sdc:
SG_IO: bad/missing sense data, sb[]:  f0 00 01 00 50 40 81 0a 00 81 00 00 00 1d 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
 drive state is:  unknown

while sdparm does not list SCT or STANDBY

sdparm -l -a /dev/sdb
    /dev/sdb: ATA       ST10000NE0008-1Z  SS02
    Direct access device specific parameters: WP=0  DPOFUA=0
Read write error recovery [rw] mode page:
  AWRE          1  [cha: n, def:  1]  Automatic write reallocation enabled
  ARRE          0  [cha: n, def:  0]  Automatic read reallocation enabled
  TB            0  [cha: n, def:  0]  Transfer block
  RC            0  [cha: n, def:  0]  Read continuous
  EER           0  [cha: n, def:  0]  Enable early recovery (obsolete)
  PER           0  [cha: n, def:  0]  Post error
  DTE           0  [cha: n, def:  0]  Data terminate on error
  DCR           0  [cha: n, def:  0]  Disable correction (obsolete)
  RRC           0  [cha: n, def:  0]  Read retry count
  COR_S         0  [cha: n, def:  0]  Correction span (obsolete)
  HOC           0  [cha: n, def:  0]  Head offset count (obsolete)
  DSOC          0  [cha: n, def:  0]  Data strobe offset count (obsolete)
  LBPERE        0  [cha: n, def:  0]  Logical block provisioning error reporting enabled
  MWR           0  [cha: n, def:  0]  Misaligned write reporting
  WRC           0  [cha: n, def:  0]  Write retry count
  RTL           0  [cha: n, def:  0]  Recovery time limit (ms)
Caching (SBC) [ca] mode page:
  IC            0  [cha: n, def:  0]  Initiator control
  ABPF          0  [cha: n, def:  0]  Abort pre-fetch
  CAP           0  [cha: n, def:  0]  Caching analysis permitted
  DISC          0  [cha: n, def:  0]  Discontinuity
  SIZE          0  [cha: n, def:  0]  Size enable
  WCE           1  [cha: y, def:  1]  Write cache enable
  MF            0  [cha: n, def:  0]  Multiplication factor
  RCD           0  [cha: n, def:  0]  Read cache disable
  DRRP          0  [cha: n, def:  0]  Demand read retention priority
  WRP           0  [cha: n, def:  0]  Write retention priority
  DPTL          0  [cha: n, def:  0]  Disable pre-fetch transfer length
  MIPF          0  [cha: n, def:  0]  Minimum pre-fetch
  MAPF          0  [cha: n, def:  0]  Maximum pre-fetch
  MAPFC         0  [cha: n, def:  0]  Maximum pre-fetch ceiling
  FSW           0  [cha: n, def:  0]  Force sequential write
  LBCSS         0  [cha: n, def:  0]  Logical block cache segment size
  DRA           0  [cha: n, def:  0]  Disable read ahead
  SYNC_PROG     0  [cha: n, def:  0]  Synchronous cache progress indication
  NV_DIS        0  [cha: n, def:  0]  Non-volatile cache disable
  NCS           0  [cha: n, def:  0]  Number of cache segments
  CSS           0  [cha: n, def:  0]  Cache segment size
Control [co] mode page:
  TST           0  [cha: n, def:  0]  Task set type
  TMF_ONLY      0  [cha: n, def:  0]  Task management functions only
  DPICZ         0  [cha: n, def:  0]  Disable protection information check if protect field zero
  D_SENSE       0  [cha: y, def:  0]  Descriptor format sense data
  GLTSD         1  [cha: n, def:  1]  Global logging target save disable
  RLEC          0  [cha: n, def:  0]  Report log exception condition
  QAM           0  [cha: n, def:  0]  Queue algorithm modifier
  NUAR          0  [cha: n, def:  0]  No unit attention on release
  QERR          0  [cha: n, def:  0]  Queue error management
  VS_CTL        0  [cha: n, def:  0]  Vendor specific [byte 4, bit 7]
  RAC           0  [cha: n, def:  0]  Report a check
  UA_INTLCK     0  [cha: n, def:  0]  Unit attention interlocks control
  SWP           0  [cha: n, def:  0]  Software write protect
  ATO           0  [cha: n, def:  0]  Application tag owner
  TAS           0  [cha: n, def:  0]  Task aborted status
  ATMPE         0  [cha: n, def:  0]  Application tag mode page enabled
  RWWP          0  [cha: n, def:  0]  Reject write without protection
  SBLP          0  [cha: n, def:  0]  Supported block lengths and protection information
  AUTOLOAD      0  [cha: n, def:  0]  Autoload mode
  BTP           -1  [cha: n, def: -1]  Busy timeout period (100us)
  ESTCT         30  [cha: n, def: 30]  Extended self test completion time (sec)

gromit · 2024-08-11 09:41:49

Don't necrobump posts please

This might be an instance of a known kernel bug that I have reported here:
https://lore.kernel.org/all/0bf3f2f0-0f … @heusel.eu

If downgrading the kernel helps it's most likely the same bug

Kardell · 2024-08-11 14:11:16

gromit wrote:

Don't necrobump posts please
This might be an instance of a known kernel bug that I have reported here:
https://lore.kernel.org/all/0bf3f2f0-0f … @heusel.eu
If downgrading the kernel helps it's most likely the same bug

make sense, I am on latest stable: 6.10.3-arch1-2.

cryptearth · 2024-08-11 14:21:48

@gromit
I don't see a necrobump here - it's still relevant to the same topic - just a very late reply ... anyway

@OP
PUIS is set with

hdparm -s1 <device>

usually you will get this warning:

[main@main ~]$ sudo hdparm -s1 /dev/sda

/dev/sda:
Use of -s1 is VERY DANGEROUS.
This requires BIOS and kernel support to recognize/boot the drive.
Please supply the --yes-i-know-what-i-am-doing flag if you really want this.
Program aborted.
[main@main ~]$

unfortunate at least I don't know of a way to get any information about if the controller or the drive does support it other than just yolo and try it
for me when I set my drives to PUIS they will get started one at a time very early on boot when udev runs the device discovery - but I know that this is specific to how my hardware and drives works
other possible options:
- neither the drives nor the controller does support it anyway - tryin to use -s1 to enable PUIS will do nothin
- only one of either the controller of the drives support it but not both - maybe same result as above - but could also lead to "dead" drives
- both controller and drives support it - should work fine as long as you controller is also support by the kernel
- any other combination or additional drivers or driver configuration options or option-rom / firmware of the controller - anything between nothin, "dead" drives and working

!be aware - you've been warned - this can lead to non-working drives with incompatible hardware - do it on your own risk!

the s-ata specs has another feature called staggered spin-up - from the effect it's the same: the drive isn't spun up until the controller tells it to - but it requires a special power connector often only found on backplanes require special controllers and sometimes software - maybe a special-purpose NAS does provide such a feature

anyway - when using an array for occasional backups I would look for an external enclosure without custom frmware but a regular amd64 hardware and setup my own system I then would only power up for backups and after checking the backup is good power back down - this also prevents you from losing an always mounted backup by some ransomware (which cause a lot of data loss for many who did that)

Last edited by cryptearth (2024-08-11 14:23:19)

Arch Linux

#1 2024-05-30 23:58:11

[SOLVED] Daily backups - spinning disks consideration

#2 2024-05-31 00:04:29

Re: [SOLVED] Daily backups - spinning disks consideration

#3 2024-05-31 00:14:16

Re: [SOLVED] Daily backups - spinning disks consideration

#4 2024-05-31 05:45:30

Re: [SOLVED] Daily backups - spinning disks consideration

#5 2024-08-11 03:31:04

Re: [SOLVED] Daily backups - spinning disks consideration

#6 2024-08-11 09:41:49

Re: [SOLVED] Daily backups - spinning disks consideration

#7 2024-08-11 14:11:16

Re: [SOLVED] Daily backups - spinning disks consideration

#8 2024-08-11 14:21:48

Re: [SOLVED] Daily backups - spinning disks consideration

Board footer