SOLVED Upgrading existing system to a Rack Chassis

[Sumppi]

I'm the process of adding some new HDD's to my existing FreeNAS System,
and now I'm ready to pull trigger on a rack chassis with backplane etc.

Initally, I planned to buy a bigger ATX PSU and cram the new HDD's in the current chassis.
Since I've seen the error of my way, and now I'm strongly leaning towards a proper chassis.

My current system:
MB: Supermicro X10SL7-F
CPU: Intel® Xeon® Processor E3-1246 v3
CPU cooler: ARCTIC Alpine 12 CO
RAM: Micron 32GB KIT 4X8GB PC3L-12800E DDR3-1600Mhz
HDD: 9 * 10TB Western Digital white labeled RED-drive (shucked from WD My Book) + 1TB NVME via PCI-E adapted
ZFS-Pool: 8 * 10TB in RAID-Z2 (remaining 10TB is a spare, acting as a temp storagate) - Connected to the onboard HBA
PSU: Seasonic Focus GX 650W
Fans: number of ARCTIC P14 PWM PST and ARCTIC P12 PWM PST


Planned extension (existing drives will remain in the system):
HDD: 8 * 14TB Western Digital White label RED-drive (Shucked from WD elements) - Already bought
ZFS-Pool: 8 * 14TB in RAID-Z2
HBA: LSI 6Gbps SAS HBA LSI 9211-8 - Already bought from eBay
Chassis: SuperChassis 846E16-R1200B in eBay
Expander backplane: SAS2 Expander Backplane BPN-SAS2-846EL1
PSU: Noise will be an issue so the included PSU's will be swapped to Supermicro 920W 1U Redundant Power Supply (PWS-920P-SQ)
I guess it's a good idea to go for redundant PSU's?
FANS: Existing fans will be replaced by a necessary number of static pressure optimized fans, ARCTIC P14 PWM PST, ARCTIC P12 PWM PST and Artic P8 PWM PST

The server will be located in my workroom/study, so noice is somewhat of an issue.
So far, I've been using the excellent Script: Hybrid CPU & HD Fan Zone Controller to keep HDD's cool at all loads (below 35C).

The excelent SAS primer gave some confidence, that I should be on the correct path.
Though I'm still a bit unsure, if I can utilize dual-links for the backplane. A single link should suffice for my use-cases (Plex Media Server).

How does my plan seems?
At least cooling might be an issue in a rack chassis. If this is the case, I will prioritize cooling over the noice, I don't want to bake my expensive hardware.

The burn-in process for the new HDD's and HBA will be done before taking the chassis/backplane in use.
I was planning to use solnet-array-test to burn-in the backplane, and only using new HDD's. Should I disconnect the exising pool for this phase?

 

[sretalla]

Though I'm still a bit unsure, if I can utilize dual-links for the backplane

Depends on the backplane, but it's possible:

A single link should suffice for my use-cases (Plex Media Server)

But that's also true.

I was planning to use solnet-array-test to burn-in the backplane, and only using new HDD's. Should I disconnect the exising pool for this phase?

In order to remove all risk of accidentally hitting pool disks with burn in routines, that would be a good move.

I guess it's a good idea to go for redundant PSU's?

If you have a good UPS in front of your system and can afford a few hours of downtime every 5 or so years to replace the power supply, I don't think it's essential (as you said, it's just Plex), but it's an OK thing to do.

 

[Jailer]

Unless I'm mistaken you should be able to utilize the onboard SAS controller with the proper cables and skip buying another HBA.

 

[Sumppi]

Depends on the backplane, but it's possible:

Yeah, that's what I've gathered. The expander backplane (SAS2 Expander Backplane BPN-SAS2-846EL1) contains three SAS-ports, one of which is for downstream expander, and using the remaining two for a single HBA is the open question.

In order to remove all risk of accidentally hitting pool disks with burn in routines, that would be a good move.

Yeah, I'll do just that.

If you have a good UPS in front of your system and can afford a few hours of downtime every 5 or so years to replace the power supply, I don't think it's essential (as you said, it's just Plex), but it's an OK thing to do.

No UPS yet, but that's something I'm planning on getting at some point.
Redundant PSU would also give me some headroom for future expansions.

Unless I'm mistaken you should be able to utilize the onboard SAS controller with the proper cables and skip buying another HBA.

No, you're not mistaken. I could utilize the onboard controller with reverse breakout cables.
However, I had already bought the HBA before this use-case for relevant.

 

[sretalla]

Redundant PSU would also give me some headroom for future expansions.

Well, not really... additional power when both supplies are on is fine, but doing the planning right, you need to allow for the whole system to be able to run on a single one or it's not really redundant (failure of one will take the whole system down if the remaining one can't cope).

 

[Sumppi]

Well, not really... additional power when both supplies are on is fine, but doing the planning right, you need to allow for the whole system to be able to run on a single one or it's not really redundant (failure of one will take the whole system down if the remaining one can't cope).

Yeah, that's true. However, I'm not planning on exceeding what a single PSU can handle. When a single can barely handle the load, two can handle it with ease. I don't want to over-stress the PSU.

 

[HoneyBadger]

Yeah, that's true. However, I'm not planning on exceeding what a single PSU can handle. When a single can barely handle the load, two can handle it with ease. I don't want to over-stress the PSU.

The largest power draw you'll see on these systems is typically during power-on, and that would only be if you were spinning up all drives at once; and most backplane/SAS controller combos don't do this, instead staggering spin-up to avoid that big surge of inrush current.

A single PSU has to be able to bring that system up from a dead state, otherwise as posited by @sretalla it's not really redundancy. As long as the 920W is certified for the chassis, a single one will be fine. the secondary is great for a spare but my priority would likely be a UPS first, and then a second PSU (on a second, separate UPS, or at least on a separate circuit to your breaker)

 

[Sumppi]

The largest power draw you'll see on these systems is typically during power-on, and that would only be if you were spinning up all drives at once; and most backplane/SAS controller combos don't do this, instead staggering spin-up to avoid that big surge of inrush current.

Yeah, the startup is what I had mind with this.
At least Proper Power Supply Sizing Guidance in general recommends against staggered spinup.
Backplanes do not do this without configuration?

A single PSU has to be able to bring that system up from a dead state, otherwise as posited by @sretalla it's not really redundancy. As long as the 920W is certified for the chassis, a single one will be fine. the secondary is great for a spare but my priority would likely be a UPS first, and then a second PSU (on a second, separate UPS, or at least on a separate circuit to your breaker)

I checked, the 920W is certified and validated for this chassis.
Thanks for the input, I'll be prioritizing a UPS!

But anyhow, the system in general is looking good?

 

[HoneyBadger]

Yeah, the startup is what I had mind with this.
At least Proper Power Supply Sizing Guidance in general recommends against staggered spinup.
Backplanes do not do this without configuration?

If you're referencing the part where it says it's "fraught with peril" note that the statement is under the heading of "Aww, do I have to?" which is implying that one hasn't purchased a sufficiently beefy PSU and is instead trying to use a weak one (eg: 350W for 8-12 drives) and is counting on staggered spinup to save them from blowing a fuse. As suggested, it often doesn't work out as planned.

Regular server iron will frequently do staggered spinup by default.

I checked, the 920W is certified and validated for this chassis.
Thanks for the input, I'll be prioritizing a UPS!

But anyhow, the system in general is looking good?

Provided that you were happy with the performance before, it looks good. What's the 1TB NVMe drive being used for?

 

[Sumppi]

If you're referencing the part where it says it's "fraught with peril" note that the statement is under the heading of "Aww, do I have to?" which is implying that one hasn't purchased a sufficiently beefy PSU and is instead trying to use a weak one (eg: 350W for 8-12 drives) and is counting on staggered spinup to save them from blowing a fuse. As suggested, it often doesn't work out as planned.

Regular server iron will frequently do staggered spinup by default.

Ok, thanks for the info!

Provided that you were happy with the performance before, it looks good. What's the 1TB NVMe drive being used for?

I have been very happy with the performance of my NAS. IIRC read/write speeds have been around 450 to 550MB/s. NVME is mainly for Plex metadata (with regural backups), and temporary storage for a torrent client (download folder).

 

[Sumppi]

The burn-in process for the new HDD's and backplane has been successfully completed.

The main caveat from the burn-in process: Cooling solution.
Current cooling can keep the drives cool enough (approx. 37C at load), but that requires near a 100% duty-cycle (PWM-controlled fans).
I find this insufficient as the ambient temperature will be a quite a bit higher in the summer months.
Fans will be replaced with Noctua:
- 3 * NF-F12 INDUSTRIALPPC-2000 PWM (mounted in the fan-wall of the chassis)
- 3 * NF-A14 INDUSTRIALPPC-2000 PWM (externally mounted with 3D-printed fan bezel - this model is somewhat broken, so I drew my own).

Now that HW is ready, I'm still pondering about pool layout. The current pool has a single VDev (8 * 10TB in RAIDZ2).
The pool is quite full (about 20% of space left).

Options are:
1) Create a new pool from the new drives, in a single VDev with 8 * 14TB in RAIDZ2
2) Extend the existing pool with a new VDev (8 * 14TB in RAIDZ).

The pool is mainly used for Plex media storage.

 

[HoneyBadger]

Given that you're using equal redundancy (RAIDZ2) in both setups, it's mostly down to whether or not you want a logical separation between the two. Extending the pool with a new vdev will not redistribute the existing data in any way, and the old "nearly-full" vdev may become a bit of a slowdown at some point for new data, but a Plex media pool doesn't exactly require a write-intensive configuration.

I'd personally make a new pool out of the 14TBs and add it as a second media source/repository in Plex.

 

[Sumppi]

Went with separate pools.

Those high performance Noctua fans are working great with the FAN control script by @Glorious1.

Thanks for all the suggestions and feedback!

 

[Sumppi]

Seems that I declared the burn-in process a success to soon.
I was performing the very last test with solnet array tester, when a single drive started acting up.
Well, at-least email notifications work nicely when something like this happens.

ZFS is yet to kick the drive out of the pool, but I suppose that's just a matter of time.
Now i'm kinda regretting that I didn't buy a spare.

Code:

smartctl 7.0 2018-12-30 r4883 [FreeBSD 11.3-RELEASE-p5 amd64] (local build)
Copyright (C) 2002-18, Bruce Allen, Christian Franke, www.smartmontools.org

=== START OF INFORMATION SECTION ===
Device Model:     WDC WD140EDFZ-11A0VA0
Serial Number:    *********
LU WWN Device Id: * ****** ******
Firmware Version: 81.00A81
User Capacity:    14,000,519,643,136 bytes [14.0 TB]
Sector Sizes:     512 bytes logical, 4096 bytes physical
Rotation Rate:    5400 rpm
Form Factor:      3.5 inches
Device is:        Not in smartctl database [for details use: -P showall]
ATA Version is:   ACS-2, ATA8-ACS T13/1699-D revision 4
SATA Version is:  SATA 3.2, 6.0 Gb/s (current: 6.0 Gb/s)
Local Time is:    Sat Dec 19 18:40:53 2020 EET
SMART support is: Available - device has SMART capability.
SMART support is: Enabled

=== START OF READ SMART DATA SECTION ===
SMART overall-health self-assessment test result: PASSED

General SMART Values:
Offline data collection status:  (0x80)    Offline data collection activity
                    was never started.
                    Auto Offline Data Collection: Enabled.
Self-test execution status:      (   0)    The previous self-test routine completed
                    without error or no self-test has ever
                    been run.
Total time to complete Offline
data collection:         (  101) seconds.
Offline data collection
capabilities:              (0x5b) SMART execute Offline immediate.
                    Auto Offline data collection on/off support.
                    Suspend Offline collection upon new
                    command.
                    Offline surface scan supported.
                    Self-test supported.
                    No Conveyance Self-test supported.
                    Selective Self-test supported.
SMART capabilities:            (0x0003)    Saves SMART data before entering
                    power-saving mode.
                    Supports SMART auto save timer.
Error logging capability:        (0x01)    Error logging supported.
                    General Purpose Logging supported.
Short self-test routine
recommended polling time:      (   2) minutes.
Extended self-test routine
recommended polling time:      (1402) minutes.
SCT capabilities:            (0x003d)    SCT Status supported.
                    SCT Error Recovery Control supported.
                    SCT Feature Control supported.
                    SCT Data Table supported.

SMART Attributes Data Structure revision number: 16
Vendor Specific SMART Attributes with Thresholds:
ID# ATTRIBUTE_NAME          FLAG     VALUE WORST THRESH TYPE      UPDATED  WHEN_FAILED RAW_VALUE
  1 Raw_Read_Error_Rate     0x000b   100   100   001    Pre-fail  Always       -       131072
  2 Throughput_Performance  0x0004   135   135   054    Old_age   Offline      -       104
  3 Spin_Up_Time            0x0007   082   082   001    Pre-fail  Always       -       388 (Average 363)
  4 Start_Stop_Count        0x0012   100   100   000    Old_age   Always       -       24
  5 Reallocated_Sector_Ct   0x0033   100   100   001    Pre-fail  Always       -       0
  7 Seek_Error_Rate         0x000a   100   100   001    Old_age   Always       -       0
  8 Seek_Time_Performance   0x0004   133   133   020    Old_age   Offline      -       18
  9 Power_On_Hours          0x0012   100   100   000    Old_age   Always       -       390
10 Spin_Retry_Count        0x0012   100   100   001    Old_age   Always       -       0
12 Power_Cycle_Count       0x0032   100   100   000    Old_age   Always       -       24
22 Unknown_Attribute       0x0023   100   100   025    Pre-fail  Always       -       100
192 Power-Off_Retract_Count 0x0032   100   100   000    Old_age   Always       -       33
193 Load_Cycle_Count        0x0012   100   100   000    Old_age   Always       -       33
194 Temperature_Celsius     0x0002   048   048   000    Old_age   Always       -       34 (Min/Max 23/38)
196 Reallocated_Event_Count 0x0032   100   100   000    Old_age   Always       -       0
197 Current_Pending_Sector  0x0022   100   100   000    Old_age   Always       -       8
198 Offline_Uncorrectable   0x0008   100   100   000    Old_age   Offline      -       0
199 UDMA_CRC_Error_Count    0x000a   100   100   000    Old_age   Always       -       0

SMART Error Log Version: 1
ATA Error Count: 1
    CR = Command Register [HEX]
    FR = Features Register [HEX]
    SC = Sector Count Register [HEX]
    SN = Sector Number Register [HEX]
    CL = Cylinder Low Register [HEX]
    CH = Cylinder High Register [HEX]
    DH = Device/Head Register [HEX]
    DC = Device Command Register [HEX]
    ER = Error register [HEX]
    ST = Status register [HEX]
Powered_Up_Time is measured from power on, and printed as
DDd+hh:mm:SS.sss where DD=days, hh=hours, mm=minutes,
SS=sec, and sss=millisec. It "wraps" after 49.710 days.

Error 1 occurred at disk power-on lifetime: 390 hours (16 days + 6 hours)
  When the command that caused the error occurred, the device was active or idle.

  After command completion occurred, registers were:
  ER ST SC SN CL CH DH
  -- -- -- -- -- -- --
  40 41 00 00 00 00 00  Error: UNC at LBA = 0x00000000 = 0

  Commands leading to the command that caused the error were:
  CR FR SC SN CL CH DH DC   Powered_Up_Time  Command/Feature_Name
  -- -- -- -- -- -- -- --  ----------------  --------------------
  60 00 00 00 43 45 40 00      19:45:13.001  READ FPDMA QUEUED
  60 00 00 00 42 45 40 00      19:45:10.266  READ FPDMA QUEUED
  60 00 00 00 41 45 40 00      19:45:10.266  READ FPDMA QUEUED
  60 00 00 00 40 45 40 00      19:45:10.260  READ FPDMA QUEUED
  60 00 00 00 3f 45 40 00      19:45:10.260  READ FPDMA QUEUED

SMART Self-test log structure revision number 1
Num  Test_Description    Status                  Remaining  LifeTime(hours)  LBA_of_first_error
# 1  Short offline       Completed: read failure       90%       390         423969760
# 2  Short offline       Completed without error       00%       375         -
# 3  Extended offline    Completed without error       00%       229         -
# 4  Extended offline    Completed without error       00%        22         -
# 5  Short offline       Completed without error       00%         0         -

SMART Selective self-test log data structure revision number 1
SPAN  MIN_LBA  MAX_LBA  CURRENT_TEST_STATUS
    1        0        0  Not_testing
    2        0        0  Not_testing
    3        0        0  Not_testing
    4        0        0  Not_testing
    5        0        0  Not_testing
Selective self-test flags (0x0):
  After scanning selected spans, do NOT read-scan remainder of disk.
If Selective self-test is pending on power-up, resume after 0 minute delay.

*EDIT* Amazon RMAed the drive in question, a replacement is on the way.