BRAND NEW HGST 4TB SAS HUS726040ALS214 for under $70 ($55 for qty >= 10)

Notice: Page may contain affiliate links for which we may earn a small commission through services like Amazon Affiliates or Skimlinks.

BLinux

cat lover server enthusiast
Jul 7, 2016
2,669
1,081
113
artofserver.com
Update. My third drive replacement (& resilvering) was a DOA. It spins up, but makes high pitch noises. I let it be to see, but after the resilvering my pool was degraded because of too many checksum errors.

I pulled the drive and put in a new one. I hope that was the only one (since I only ordered one extra).

I guess I should contact the seller to see what they want to do.
I believe Scott will take care of you...

The irony is that your drives were very well packaged with thick foam protection and yet you got a dud. My 1st two orders were also in the same packaging and i got 2 duds. My 3rd order came in what i would consider lesser packaging, and all of those were good.
 
  • Like
Reactions: nle

nle

Member
Oct 24, 2012
204
11
18
Yes, I have a good impression as well.

I will try to complete the exchange and resilvering of drives, should only take a couple of more days if all is good with the rest, before I contact him. The only issue I see is that the postage could potentially cost more than it's worth. We'll see.
 

nle

Member
Oct 24, 2012
204
11
18
So I might have another (or possibly two) dud. I have completed the exchange and resilvered my entire pool, and is currently running long S.M.A.R.T tests. (I only have a single host with SAS capability [Dell R710])

One drive is making weird sounds
Not sure which one, so I guess I need to pull each one until I find it. Just need to shut down my services.

One drive refuses to run a self test, and is also showing some errors
Hopefully this is the same drive that are making the sounds.
Code:
smartctl 6.6 2017-11-05 r4594 [i386-pc-solaris2.11] (local build)
Copyright (C) 2002-17, Bruce Allen, Christian Franke, www.smartmontools.org

=== START OF INFORMATION SECTION ===
Vendor:               HGST
Product:              HUS726040ALS214
Revision:             MS00
Compliance:           SPC-4
User Capacity:        4,000,787,030,016 bytes [4.00 TB]
Logical block size:   512 bytes
Formatted with type 2 protection
LU is fully provisioned
Rotation Rate:        7200 rpm
Form Factor:          3.5 inches
Logical Unit id:      0x5000cca25d522048
Serial number:        K4HG54ZB
Device type:          disk
Transport protocol:   SAS (SPL-3)
Local Time is:        Fri Jun  1 10:47:16 2018 CEST
SMART support is:     Available - device has SMART capability.
SMART support is:     Enabled
Temperature Warning:  Enabled

=== START OF READ SMART DATA SECTION ===
SMART Health Status: OK

Current Drive Temperature:     37 C
Drive Trip Temperature:        55 C

Manufactured in week 51 of year 2016
Specified cycle count over device lifetime:  50000
Accumulated start-stop cycles:  3
Specified load-unload count over device lifetime:  600000
Accumulated load-unload cycles:  221
Elements in grown defect list: 127

Vendor (Seagate) cache information
  Blocks sent to initiator = 319776001884160

Error counter log:
           Errors Corrected by           Total   Correction     Gigabytes    Total
               ECC          rereads/    errors   algorithm      processed    uncorrected
           fast | delayed   rewrites  corrected  invocations   [10^9 bytes]  errors
read:          0        0         0         0       3285         18.004           0
write:         0      312         0       312      25427       1441.546          37
verify:        0        0         0         0        757          0.000           0

Non-medium error count:        0

SMART Self-test log
Num  Test              Status                 segment  LifeTime  LBA_first_err [SK ASC ASQ]
     Description                              number   (hours)
# 1  Background short  Failed in segment -->       3      15                 - [-   -    -]
# 2  Background long   Failed in segment -->       3      14                 - [-   -    -]
# 3  Background long   Failed in segment -->       3      13                 - [-   -    -]

Long (extended) Self Test duration: 6 seconds [0.1 minutes]
My zpool is not complaining, but I have not tested writing a lot of data yet either.

@BLinux Any input on this? :)
 

BLinux

cat lover server enthusiast
Jul 7, 2016
2,669
1,081
113
artofserver.com
@nle both of the drives that I had that were duds failed SMART short/long tests multiple times. 1 of them had a few dozen of errors in both read/write and badblocks reported the same. So, yeah, unfortunately, doesn't sound good for you...
 
  • Like
Reactions: nle

BLinux

cat lover server enthusiast
Jul 7, 2016
2,669
1,081
113
artofserver.com
@nle if I recall, you're outside the USA? so, i guess returns/replacements might be difficult for you... in that case, one thing I would try is running badblocks on the drive. at least, it will attempt to write to all sectors, and maybe that will force a reallocation and make things work again (run 2nd pass of badblocks to confirm 0 errors). i don't usually have a lot of confidence in a drive that shows problems like this brand new, but some times they have a few errors, and then are stable for many years. at least, i've had several drives that were like that. if the 2nd pass of badblocks shows new errors, then definitely that drive is a goner.
 
  • Like
Reactions: nle

AFisher

Member
Jun 2, 2017
51
15
8
I reached out to Scott, and set up an order for 20 drives ($50 each) they arrived today... now my question how are you guys testing these? and how are you getting a report on that. In freeNas, I can run smartctl -a and see a report, but no txt\file as a report and I have to do it one at a time... I'ld love to run a in-depth test and get a status for each drive at the end. oh and secondly I have several drives, that freenas wont wipe\format seems there is a gpt partition I can't seem to get removed, any I deas? I try <DD if=/dev/zero of=da## bs=512 count=1> where ## is a number corresponding to the drive
 

BLinux

cat lover server enthusiast
Jul 7, 2016
2,669
1,081
113
artofserver.com
I reached out to Scott, and set up an order for 20 drives ($50 each) they arrived today... now my question how are you guys testing these? and how are you getting a report on that. In freeNas, I can run smartctl -a and see a report, but no txt\file as a report and I have to do it one at a time... I'ld love to run a in-depth test and get a status for each drive at the end. oh and secondly I have several drives, that freenas wont wipe\format seems there is a gpt partition I can't seem to get removed, any I deas? I try <DD if=/dev/zero of=da## bs=512 count=1> where ## is a number corresponding to the drive
I wrote a script to run a SMART short test, then run badblocks, and then run SMART long test and save the results output to a file. I run badblocks inside "screen", so if I'm curious, I can pop in to see how it is going. I can also recursively run tail -1 on the output files to get quick glance on the testing progress. I have a 24-bay Supermicro 846 and the script will launch all 24 (or however many I've configured it for) tests in parallel. For these 4TB drives, it takes about 60 hours to complete the test, some finish earlier than others, but that's about how long.

FreeNAS is very particular. Hard drives that worked fine with ZFS on Linux, may not work in FreeNAS properly and their error messages are not helpful. What I've found that works to resolve this type of issue in the past is to wipe the entire drive with "0" or run a low level format with the sg_format tools. The process does take some time though.. seems like such a waste of energy and time; it truly is something i find very annoying about FreeNAS.
 

AFisher

Member
Jun 2, 2017
51
15
8
I wrote a script to run a SMART short test, then run badblocks, and then run SMART long test and save the results output to a file. I run badblocks inside "screen", so if I'm curious, I can pop in to see how it is going. I can also recursively run tail -1 on the output files to get quick glance on the testing progress. I have a 24-bay Supermicro 846 and the script will launch all 24 (or however many I've configured it for) tests in parallel. For these 4TB drives, it takes about 60 hours to complete the test, some finish earlier than others, but that's about how long.

FreeNAS is very particular. Hard drives that worked fine with ZFS on Linux, may not work in FreeNAS properly and their error messages are not helpful. What I've found that works to resolve this type of issue in the past is to wipe the entire drive with "0" or run a low level format with the sg_format tools. The process does take some time though.. seems like such a waste of energy and time; it truly is something i find very annoying about FreeNAS.
any chance you'd be will to share said script? Ive been try to get something similar together, but my bash is weak!
 
  • Like
Reactions: Patriot

svtkobra7

Active Member
Jan 2, 2017
362
87
28
any chance you'd be will to share said script? Ive been try to get something similar together, but my bash is weak!
Take a look at this, it is what I used: Spearfoot/disk-burnin-and-testing

Sounds very similar to the aforementioned script and accomplishes the following (with logging):
  • Run SMART short test
  • Run SMART extended test
  • Run badblocks
  • Run SMART short test
  • Run SMART extended test
Code:
#!/bin/sh
########################################################################
#
# disk-burnin.sh
#
# A script to simplify the process of burning-in disks. Intended for use
# only on disks which do not contain valuable data, such as new disks or
# disks which are being tested or re-purposed.
#
# Be aware that:
#
#   1> This script runs the badblocks program in destructive mode, which
#      erases any data on the disk.
#
#   !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
#   !!!!!        WILL DESTROY THE DISK CONTENTS! BE CAREFUL!        !!!!!
#   !!!!! DO NOT RUN THIS SCRIPT ON DISKS CONTAINING DATA YOU VALUE !!!!!
#   !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
#
#   2> Run times for large disks can take several days to complete, so it
#      is a good idea to use tmux sessions to prevent mishaps.
#
#   3> Must be run as 'root'.
#
#   4> Tests of large drives can take days to complete: use tmux!
#
# Performs these steps:
#
#   1> Run SMART short test
#   2> Run SMART extended test
#   3> Run badblocks
#   4> Run SMART short test
#   5> Run SMART extended test
#
# The script sleeps after starting each SMART test, using a duration
# based on the polling interval reported by the disk, after which the
# script will poll the disk to verify the self-test has completed.
#
# Full SMART information is pulled after each SMART test. All output
# except for the sleep command is echoed to both the screen and log file.
#
# You should monitor the burn-in progress and watch for errors, particularly
# any errors reported by badblocks, or these SMART errors:
#
#   5 Reallocated_Sector_Ct   
# 196 Reallocated_Event_Count
# 197 Current_Pending_Sector 
# 198 Offline_Uncorrectable   
#
# These indicate possible problems with the drive. You therefore may
# wish to abort the remaining tests and proceed with an RMA exchange
# for new drives or discard old ones. Also please note that this list
# is not exhaustive.
#
# The script extracts the drive model and serial number and forms
# a log filename of the form 'burnin-[model]_[serial number].log'.
#
# badblocks is invoked with a block size of 4096, the -wsv options, and
# the -o option to instruct it to write the list of bad blocks found (if
# any) to a file named 'burnin-[model]_[serial number].bb'.
#
# The only required command-line argument is the device specifier, e.g.:
#
#   ./disk-burnin.sh sda
#
# ...will run the burn-in test on device /dev/sda
#
# You can run the script in 'dry run mode' (see below) to check the sleep
# duration calculations and to insure that the sequence of commands suits
# your needs. In 'dry runs' the script does not actually perform any
# SMART tests or invoke the sleep or badblocks programs. The script is
# distributed with 'dry runs' enabled, so you will need to edit the
# Dry_Run variable below, setting it to 0, in order to actually perform
# tests on drives.
#
# Before using the script on FreeBSD systems (including FreeNAS) you must
# first execute this sysctl command to alter the kernel's geometry debug
# flags. This allows badblocks to write to the entire disk:
#
#   sysctl kern.geom.debugflags=0x10
#
# Tested under:
#   FreeNAS 9.10.2 (FreeBSD 10.3-STABLE)
#   Ubuntu Server 16.04.2 LTS
#
# Tested on:
#   Intel DC S3700 SSD
#   Intel Model 320 Series SSD
#   HGST Deskstar NAS (HDN724040ALE640)
#   Hitachi/HGST Ultrastar 7K4000 (HUS724020ALE640)
#   Western Digital Re (WD4000FYYZ)
#   Western Digital Black (WD6001FZWX)
#
# Requires the smartmontools, available at https://www.smartmontools.org
#
# Uses: grep, pcregrep, awk, sed, tr, sleep, badblocks
#
# Written by Keith Nash, March 2017
#
# KN, 8 Apr 2017:
#   Added minimum test durations because some devices don't return accurate values.
#   Added code to clean up the log file, removing copyright notices, etc.
#   No longer echo 'smartctl -t' output to log file as it imparts no useful information.
#   Emit test results after tests instead of full 'smartctl -a' output.
#   Emit full 'smartctl -x' output at the end of all testing.
#   Minor changes to log output and formatting.
#
# KN, 12 May 2017:
#   Added code to poll the disk and check for completed self-tests.
#
#   As noted above, some disks don't report accurate values for the short and extended
#   self-test intervals, sometimes by a significant amount. The original approach using
#   'fudge' factors wasn't reliable and the script would finish even though the SMART
#   self-tests had not completed. The new polling code helps insure that this doesn't
#   happen.
#   
#   Fixed code to work around annoying differences between sed's behavior on Linux and
#   FreeBSD.
#
# KN, 8 Jun 2017
#   Modified parsing of short and extended test durations to accommodate the values
#   returned by larger drives; we needed to strip out the '(' and ')' characters
#   surrounding the integer value in order to fetch it reliably.
#
########################################################################

if [ $# -ne 1 ]; then
  echo "Error: not enough arguments!"
  echo "Usage is: $0 drive_device_specifier"
  exit 2
fi

Drive=$1

# Set Dry_Run to a non-zero value to test out the script without actually
# running any tests: set it to zero when you are ready to burn-in disks.

Dry_Run=1

# Directory specifiers for log and badblocks data files. Leave off the
# trailing slash:

Log_Dir="."
BB_Dir="."

########################################################################
#
# Prologue
#
########################################################################

# Obtain the disk model and serial number:

Disk_Model=$(smartctl -i /dev/"$Drive" | grep "Device Model" | awk '{print $3, $4, $5}' | sed -e 's/^[ \t]*//;s/[ \t]*$//' | sed -e 's/ /_/')

if [ -z "$Disk_Model" ]; then
  Disk_Model=$(smartctl -i /dev/"$Drive" | grep "Model Family" | awk '{print $3, $4, $5}' | sed -e 's/^[ \t]*//;s/[ \t]*$//' | sed -e 's/ /_/')
fi

Serial_Number=$(smartctl -i /dev/"$Drive" | grep "Serial Number" | awk '{print $3}' | sed -e 's/ /_/')

# Form the log and bad blocks data filenames:

Log_File="burnin-${Disk_Model}_${Serial_Number}.log"
Log_File=$Log_Dir/$Log_File

BB_File="burnin-${Disk_Model}_${Serial_Number}.bb"
BB_File=$BB_Dir/$BB_File

# Query the short and extended test duration, in minutes. Use the values to
# calculate how long we should sleep after starting the SMART tests:

Short_Test_Minutes=$(smartctl -c /dev/"$Drive" | pcregrep -M "Short self-test routine.*\n.*recommended polling time:" | sed -e 's/)//;s/(//' | awk '{print $4}' | tr -d '\n')
#printf "Short_Test_Minutes=[%s]\n" ${Short_Test_Minutes}

Extended_Test_Minutes=$(smartctl -c /dev/"$Drive" | pcregrep -M "Extended self-test routine.*\n.*recommended polling time:" | sed -e 's/)//;s/(//' | awk '{print $4}' | tr -d '\n')
#printf "Extended_Test_Minutes=[%s]\n" ${Extended_Test_Minutes}

Short_Test_Sleep=$((Short_Test_Minutes*60))
Extended_Test_Sleep=$((Extended_Test_Minutes*60))

# Selftest polling timeout interval, in hours
Poll_Timeout_Hours=4

# Calculate the selftest polling timeout interval in seconds
Poll_Timeout=$((Poll_Timeout_Hours*60*60))

# Polling sleep interval, in seconds:
Poll_Interval=15

########################################################################
#
# Local functions
#
########################################################################

echo_str()
{
  echo "$1" | tee -a "$Log_File"
}

push_header()
{
  echo_str "+-----------------------------------------------------------------------------"
}

poll_selftest_complete()
{
  l_rv=1
  l_status=0
  l_done=0
  l_pollduration=0

# Check SMART results for "The previous self-test routine completed"
# Return 0 if the test has completed, 1 if we exceed our polling timeout interval

  while [ $l_done -eq 0 ];
  do 
    smartctl -a /dev/"$Drive" | grep -i "The previous self-test routine completed" > /dev/null 2<&1
    l_status=$?
    if [ $l_status -eq 0 ]; then
      echo_str "SMART self-test complete"
      l_rv=0
      l_done=1
    else
      # Check for failure   
      smartctl -a /dev/"$Drive" | grep -i "of the test failed." > /dev/null 2<&1
      l_status=$?
      if [ $l_status -eq 0 ]; then
        echo_str "SMART self-test failed"
        l_rv=0
        l_done=1
      else
        if [ $l_pollduration -ge "${Poll_Timeout}" ]; then
          echo_str "Timeout polling for SMART self-test status"
          l_done=1
        else
          sleep ${Poll_Interval}
          l_pollduration=$((l_pollduration+Poll_Interval))
        fi
      fi
    fi
  done

  return $l_rv
}

run_short_test()
{
  push_header
  echo_str "+ Run SMART short test on drive /dev/${Drive}: $(date)"
  push_header
  if [ "${Dry_Run}" -eq 0 ]; then
    smartctl -t short /dev/"$Drive"
    echo_str "Short test started, sleeping ${Short_Test_Sleep} seconds until it finishes"
    sleep ${Short_Test_Sleep}
    poll_selftest_complete
    smartctl -l selftest /dev/"$Drive" | tee -a "$Log_File"
  else
    echo_str "Dry run: would start the SMART short test and sleep ${Short_Test_Sleep} seconds until the test finishes"
  fi
  echo_str "Finished SMART short test on drive /dev/${Drive}: $(date)"
}

run_extended_test()
{
  push_header
  echo_str "+ Run SMART extended test on drive /dev/${Drive}: $(date)"
  push_header
  if [ "${Dry_Run}" -eq 0 ]; then
    smartctl -t long /dev/"$Drive"
    echo_str "Extended test started, sleeping ${Extended_Test_Sleep} seconds until it finishes"
    sleep ${Extended_Test_Sleep}
    poll_selftest_complete
    smartctl -l selftest /dev/"$Drive" | tee -a "$Log_File"
  else
    echo_str "Dry run: would start the SMART extended test and sleep ${Extended_Test_Sleep} seconds until the test finishes"
  fi
  echo_str "Finished SMART extended test on drive /dev/${Drive}: $(date)"
}

run_badblocks_test()
{
  push_header
  echo_str "+ Run badblocks test on drive /dev/${Drive}: $(date)"
  push_header
  if [ "${Dry_Run}" -eq 0 ]; then
#
#   This is the command which erases all data on the disk:
#
    badblocks -b 4096 -wsv -o "$BB_File" /dev/"$Drive"
  else
    echo_str "Dry run: would run badblocks -b 4096 -wsv -o ${BB_File} /dev/${Drive}"
  fi
  echo_str "Finished badblocks test on drive /dev/${Drive}: $(date)"
}

########################################################################
#
# Action begins here
#
########################################################################

if [ -e "$Log_File" ]; then
  rm "$Log_File"
fi

push_header
echo_str "+ Started burn-in of /dev/${Drive} : $(date)"
push_header

echo_str "Host: $(hostname)"
echo_str "Drive Model: ${Disk_Model}"
echo_str "Serial Number: ${Serial_Number}"
echo_str "Short test duration: ${Short_Test_Minutes} minutes"
echo_str "Short test sleep duration: ${Short_Test_Sleep} seconds"
echo_str "Extended test duration: ${Extended_Test_Minutes} minutes"
echo_str "Extended test sleep duration: ${Extended_Test_Sleep} seconds"
echo_str "Log file: ${Log_File}"
echo_str "Bad blocks file: ${BB_File}"

# Run the test sequence:
run_short_test
run_extended_test
run_badblocks_test
run_short_test
run_extended_test

# Emit full device information to log:
push_header
echo_str "+ SMART information for drive /dev/${Drive}: $(date)"
push_header
smartctl -x /dev/"$Drive" | tee -a "$Log_File"

push_header
echo_str "+ Finished burn-in of /dev/${Drive} : $(date)"
push_header

# Clean up the log file:

osflavor=$(uname)

if [ "${osflavor}" = "Linux" ]; then
  sed -i -e '/Copyright/d' "${Log_File}"
  sed -i -e '/=== START OF READ/d' "${Log_File}"
  sed -i -e '/SMART Attributes Data/d' "${Log_File}"
  sed -i -e '/Vendor Specific SMART/d' "${Log_File}"
  sed -i -e '/SMART Error Log Version/d' "${Log_File}"
fi

if [ "${osflavor}" = "FreeBSD" ]; then
  sed -i '' -e '/Copyright/d' "${Log_File}"
  sed -i '' -e '/=== START OF READ/d' "${Log_File}"
  sed -i '' -e '/SMART Attributes Data/d' "${Log_File}"
  sed -i '' -e '/Vendor Specific SMART/d' "${Log_File}"
  sed -i '' -e '/SMART Error Log Version/d' "${Log_File}"
fi
 
  • Like
Reactions: dawsonkm and tjk

ru me

Member
Jun 2, 2018
51
30
18
I ordered four drives. One never mounted. Two have reduced read/write speed, about half of normal and smart shows bad blocks and a high number of read write errors. On top of that the drives have been used for 200 days and have about 12Tb read/write. These are going to go back. Too bad the seller wasted my time. I separately ordered two more which are indeed new and seem to have had an initial burn in of three days. I will keep those if they test out O.K.
 

fohdeesha

Kaini Industries
Nov 20, 2016
2,728
3,078
113
33
fohdeesha.com
I ordered four drives. One never mounted. Two have reduced read/write speed, about half of normal and smart shows bad blocks and a high number of read write errors. On top of that the drives have been used for 200 days and have about 12Tb read/write
wtf! I was about to order 20 or more of these under the pretense that they are actually new but this doesn't sound good
 

BLinux

cat lover server enthusiast
Jul 7, 2016
2,669
1,081
113
artofserver.com
wtf! I was about to order 20 or more of these under the pretense that they are actually new but this doesn't sound good
I've ordered 40 drives so far and all were new. I don't know who "ru me" is + first time poster makes me a bit skeptical. Either way, I say reach out to Scott and ask for assurances. I don't know Scott outside this interaction, but he has taken care of my issues with the 2 duds I got out of 40 across 3 orders.
 

BLinux

cat lover server enthusiast
Jul 7, 2016
2,669
1,081
113
artofserver.com
unrelated yet related ... is no warranty par for the course for "new pulls" ???

HGST warranty lookup suggests none (from the pictured drive in the ebay listing).

HGST Verify Waranty Eligbility

The warranty issue was discussed earlier in this thread. Because these are NetApp drives, they won't have standard warranty via HGST/WDC. At this price point, my strategy is to buy extras for self warranty. That's also why I do the burn in, identify any issues now and get them replaced by the seller. If they pass burn in, the spares go into storage for when I need them.
 

svtkobra7

Active Member
Jan 2, 2017
362
87
28
The warranty issue was discussed earlier in this thread.
My bad ... I didn't bother to search as I thought I had read every post.

At this price point, my strategy is to buy extras for self warranty.
Seriously! At that price you might as well buy 2x what you need. Sound strategy, I've always favored self insurance compared to pooled anyway (unless the cost is otherwise embedded).

That's also why I do the burn in, identify any issues now and get them replaced by the seller. If they pass burn in, the spares go into storage for when I need them.
Again, sound. In this case, it is especially relevant to burn in so they can be returned if need be, but it would be a best practice to burn in a new retail drive bought off a store shelf (if retail stores still existed that is), right? Just curious for another (more knowledgeable) perspective.

...

So, I never actually bought any as I'm in the process of building another server (currently have a 16 bay 3U 836) with 16 x 6TB. It would take a 24 bay 4U 846 to match the same raw capacity with 24 x 4TB, and as amazing as this deal is, I feel like decreasing storage density would be a step in the wrong direction (for me personally).

But I've been paying close attention to this thread as I've been considering purchasing these drives the same seller has: 8TB SAS 3.5" HUH721008AL5204 HGST ULTRASTAR HE10 7.2K 12GBPS HDD NETAPP. I hate that they are 3x ($150) the cost at qty >= 10, but they are 2x the capacity & helium :rolleyes:. At least the supposedly run 4-5C cooler than 7200 RPM non-helium drives, which would mean Optimal fan profile 100% of the time.

Any thoughts on that drive & price?
 

BLinux

cat lover server enthusiast
Jul 7, 2016
2,669
1,081
113
artofserver.com
My bad ... I didn't bother to search as I thought I had read every post.
No, sorry if that response came off as RTFM; it wasn't meant that way... just pointing to previous discussion.

Seriously! At that price you might as well buy 2x what you need. Sound strategy, I've always favored self insurance compared to pooled anyway (unless the cost is otherwise embedded).


Again, sound. In this case, it is especially relevant to burn in so they can be returned if need be, but it would be a best practice to burn in a new retail drive bought off a store shelf (if retail stores still existed that is), right? Just curious for another (more knowledgeable) perspective.
I totally agree. I do burn-in on new & used drives, and I find I weed out more bad drives from new ones. Used ones tend to have been "battle tested" already, so they are either going strong or they already show signs of problems.

So, I never actually bought any as I'm in the process of building another server (currently have a 16 bay 3U 836) with 16 x 6TB. It would take a 24 bay 4U 846 to match the same raw capacity with 24 x 4TB, and as amazing as this deal is, I feel like decreasing storage density would be a step in the wrong direction (for me personally).

But I've been paying close attention to this thread as I've been considering purchasing these drives the same seller has: 8TB SAS 3.5" HUH721008AL5204 HGST ULTRASTAR HE10 7.2K 12GBPS HDD NETAPP. I hate that they are 3x ($150) the cost at qty >= 10, but they are 2x the capacity & helium :rolleyes:. At least the supposedly run 4-5C cooler than 7200 RPM non-helium drives, which would mean Optimal fan profile 100% of the time.

Any thoughts on that drive & price?
Those 8TB @ $150 is pretty decent price, not as great as the $130 from best buy, which at least has a short warranty if you keep the USB case, but the HGST spins faster at 7200RPM. either way, their $/TB isn't as good as the 4TB, but on the other hand for the same total capacity, you are spinning half the motors with the higher $/TB. HDDs use abou 4-10W each?

one thing I was thinking about, is that all these drives are 12Gbps SAS-3, so I guess if one had a SAS-3 HBA/controller, and SAS-3 expander backplane with these 12Gbps SAS HDDs, a single 4x SAS lane cable could theoretically carry 48Gbps? (=6Gbytes/sec) ... it would be a very clean way to get a lot of bandwidth???
 

svtkobra7

Active Member
Jan 2, 2017
362
87
28
No, sorry if that response came off as RTFM; it wasn't meant that way... just pointing to previous discussion.
I didn't take it that way at all. :)

Those 8TB @ $150 is pretty decent price, not as great as the $130 from best buy, which at least has a short warranty if you keep the USB case, but the HGST spins faster at 7200RPM. either way, their $/TB isn't as good as the 4TB, but on the other hand for the same total capacity, you are spinning half the motors with the higher $/TB. HDDs use abou 4-10W each?
Re rotational speed: Double edged sword ... indeed, it spins faster, but also creates more heat (as you know). From what I've read, He drives @ 7200 RPM may be on par, temp wise, with 5400 RPM drives though.

Re energy: I'm sure everyone from CA will gasp (I'd love to know what % of the user base that is), but I don't really care about HDD energy consumption (maybe I should?). If saving tress was my MO (I like trees btw), I'd be running 96TB of non-NVME SSDs. Don't think that breakeven point is in my lifetime. You are spot on regarding consumption.

HGST Data Sheet
  • Operating 9.5W
  • Idle 5.8W
  • Power consumption efficiency at Idle = 0.58 W/TB
one thing I was thinking about, is that all these drives are 12Gbps SAS-3, so I guess if one had a SAS-3 HBA/controller, and SAS-3 expander backplane with these 12Gbps SAS HDDs, a single 4x SAS lane cable could theoretically carry 48Gbps? (=6Gbytes/sec) ... it would be a very clean way to get a lot of bandwidth???
Theoretically I believe is the key word. Maybe I missed the bus (I am, as you know, hardly IT proficient), but while the bandwidth may be there, the drives won't even saturate SAS2, so I'm unsure of the relevance of SAS3 HDDs. I think a similar comment was made earlier in this thread.;)

Thanks for your comments.

EDIT: Premature posting issue (I'll have to get that looked at).
 
Last edited:

ru me

Member
Jun 2, 2018
51
30
18
@BLinux
good for you if you got new drives. In my case, the seller packaged and sealed garbage and wasted my time. Sending used drives with defects is certainly not acceptable and other people in this thread report problems as well. Not sure why you think it would make any sense for me to be lying about this. In fact, I ordered two more drives in a separate transaction which are indeed unused. They were packaged differently and had netapp brackets.
 

ru me

Member
Jun 2, 2018
51
30
18
Here is an update. Nothing changes with regard to receiving 3 broken disks out of four. However, the tool I was using to read the smart data seems to interpret the power on hours of this particular SAS drive model wrong (HDsentinel). When I received the disks I thought they had been used extensively based on SMART. Still not sure how long the disks have been used but it looks more likely that they had been tested by somebody else and rejected. That would be consistent with about 20TB read and write.
The disks are actually easy to weed out if you can read the smart data and use a tool like HDtune. Well, one disk fails to mount but the other two bad ones have significantly reduced write speeds (peak 120 and 150 MB/s vs 200). The drives also show a high write error rate. It is not the SAS HBA, the OS or the cabling, all tested.
Make of this what you want but I hope this information will be useful to some.
 

MikeWebb

Member
Jan 28, 2018
92
29
18
53
Well here goes my first post.
Bought 8
8 arrived in so-so packaging
8 DOA

Seperate systems, changed out cables etc blah blah

Lovin this guys product. gonna get my money back