Skip site navigation (1)Skip section navigation (2)
Date:      Thu, 19 Aug 2021 04:46:46 +0100
From:      Graham Perrin <grahamperrin@gmail.com>
To:        Dan Langille <dan@langille.org>
Cc:        FreeBSD questions <freebsd-questions@freebsd.org>
Subject:   Re: nvme detached
Message-ID:  <93a02a57-9f7b-473a-a63f-1c541fdd2225@gmail.com>
In-Reply-To: <a539b43d-b32e-0636-a9cb-4928c66b08d4@langille.org>
References:  <a703ce19-ea5d-48ca-8fc6-c1f1418e3131@www.fastmail.com> <3b332fd8-24be-5a2f-15a8-630edb2a7226@gmail.com> <5ff30e22-d355-4a0c-b13b-02ac709f0fbc@www.fastmail.com> <dd806dc5-86b8-a060-e919-46cb0976180d@gmail.com> <e0282e05-2a65-4fb4-9d89-c687e8a7cb98@www.fastmail.com> <a539b43d-b32e-0636-a9cb-4928c66b08d4@langille.org>

next in thread | previous in thread | raw e-mail | index | archive | help
On 17/08/2021 16:12, Dan Langille wrote:
> Dan Langille wrote:
>>
>> On Wed, Aug 4, 2021, at 2:04 PM, Graham Perrin wrote:
>>>
>>> On 04/08/2021 18:45, Dan Langille wrote:
>>>
>>>>
>>>> On Wed, Aug 4, 2021, at 1:35 PM, Graham Perrin wrote:
>>>>
>>>>>
>>>>> A normal run of StressDesk might be enough to expose a problem; I
>>>>> recently had a new drive (less than 100 hours' use) that failed
>>>>> consistently after around seven minutes of the run (before filling the
>>>>> file UFS system).
>>>>
>>>> Is that sysutils/stressdisk?
>>>
>>>
>>> Yes, sorry for the typo.
>>
>>
>> stressdisk now underway:
>>
>> Bytes read:      26649996 MByte (1633.90 MByte/s)
>> Bytes written:     235880 MByte ( 491.77 MByte/s)
>> Errors:                 0
>> Elapsed time:  4h40m0.000371674s
>>
>> FYI, it seems my newly arrived NVME is not actually new.
>>
>> * gpart shows a partition.
>> * smarctl says 4 power on hours
>> * 56 power cycles
>> * 104GB written
>>
>> https://dan.langille.org/2021/08/09/i-bought-a-new-nvme-drive-or-did-i/
>
>
>
>
> $ sudo gpart create nvd0
> $ sudo gpart create -s gpt nvd0
> $ sudo gpart add -a 4k -t freebsd-zfs nvd0
> $ sudo zpool create nvd0 /dev/nvd0p1
> $ sudo chown dvl:dvl /nvd0
> $ stressdisk run /nvd0/ -duration 1h
>
> I returned that unit and got another.
>
> 2021/08/17 15:05:28 Exiting after running for > 1h0m0s
> 2021/08/17 15:05:28
> Bytes read:       2970440 MByte (1574.04 MByte/s)
> Bytes written:     229910 MByte ( 134.56 MByte/s)
> Errors:                 0
> Elapsed time:  1h0m0.00184772s
>
> 2021/08/17 15:05:28 PASSED with no errors
>
> In that hour, it read just over 3 TB (see smartctl output below)
>
> [dvl@test /nvd0]$ zpool list nvd0
> NAME   SIZE  ALLOC   FREE  CKPOINT  EXPANDSZ   FRAG    CAP DEDUP    
> HEALTH  ALTROOT
> nvd0   232G   224G  8.25G        -         -    36%    96% 1.00x    
> ONLINE  -
>
> [dvl@test /nvd0]$ zpool list nvd0
> NAME   SIZE  ALLOC   FREE  CKPOINT  EXPANDSZ   FRAG    CAP DEDUP    
> HEALTH  ALTROOT
> nvd0   232G   224G  8.25G        -         -    36%    96% 1.00x    
> ONLINE  -
> [dvl@test /nvd0]$ sudo smartctl -a /dev/nvme0
> smartctl 7.2 2020-12-30 r5155 [FreeBSD 13.0-RELEASE-p3 amd64] (local 
> build)
> Copyright (C) 2002-20, Bruce Allen, Christian Franke, 
> www.smartmontools.org
>
> === START OF INFORMATION SECTION ===
> Model Number:                       WDC WDS250G2B0C-00PXH0
> Serial Number:                      [redacted]
> Firmware Version:                   233010WD
> PCI Vendor/Subsystem ID:            0x15b7
> IEEE OUI Identifier:                0x001b44
> Total NVM Capacity:                 250,059,350,016 [250 GB]
> Unallocated NVM Capacity:           0
> Controller ID:                      1
> NVMe Version:                       1.4
> Number of Namespaces:               1
> Namespace 1 Size/Capacity:          250,059,350,016 [250 GB]
> Namespace 1 Formatted LBA Size:     512
> Namespace 1 IEEE EUI-64:            001b44 8b4111e9c1
> Local Time is:                      Tue Aug 17 15:11:06 2021 UTC
> Firmware Updates (0x14):            2 Slots, no Reset required
> Optional Admin Commands (0x0017):   Security Format Frmw_DL Self_Test
> Optional NVM Commands (0x005f):     Comp Wr_Unc DS_Mngmt Wr_Zero 
> Sav/Sel_Feat Timestmp
> Log Page Attributes (0x1e):         Cmd_Eff_Lg Ext_Get_Lg Telmtry_Lg 
> Pers_Ev_Lg
> Maximum Data Transfer Size:         128 Pages
> Warning  Comp. Temp. Threshold:     80 Celsius
> Critical Comp. Temp. Threshold:     85 Celsius
> Namespace 1 Features (0x02):        NA_Fields
>
> Supported Power States
> St Op     Max   Active     Idle   RL RT WL WT  Ent_Lat  Ex_Lat
> 0 +     3.50W    2.00W       -    0  0  0  0        0       0
> 1 +     2.40W    1.80W       -    0  0  0  0        0       0
> 2 +     1.90W    1.50W       -    0  0  0  0        0       0
> 3 -   0.0250W       -        -    3  3  3  3     3900   11000
> 4 -   0.0050W       -        -    4  4  4  4     5000   44000
>
> Supported LBA Sizes (NSID 0x1)
> Id Fmt  Data  Metadt  Rel_Perf
> 0 +     512       0         2
> 1 -    4096       0         1
>
> === START OF SMART DATA SECTION ===
> SMART overall-health self-assessment test result: PASSED
>
> SMART/Health Information (NVMe Log 0x02)
> Critical Warning:                   0x00
> Temperature:                        37 Celsius
> Available Spare:                    100%
> Available Spare Threshold:          10%
> Percentage Used:                    0%
> Data Units Read:                    5,980,477 [3.06 TB]
> Data Units Written:                 472,529 [241 GB]
> Host Read Commands:                 23,385,657
> Host Write Commands:                1,957,158
> Controller Busy Time:               70
> Power Cycles:                       3
> Power On Hours:                     1
> Unsafe Shutdowns:                   0
> Media and Data Integrity Errors:    0
> Error Information Log Entries:      1
> Warning  Comp. Temperature Time:    0
> Critical Comp. Temperature Time:    0
>
> Error Information (NVMe Log 0x01, 16 of 256 entries)
> No Errors Logged
>
> [dvl@test /nvd0]$


Thanks

If/when you begin using the new disk, then maybe give the old disk
entirely to UFS for a run of StressDisk (or whatever you like) to tell
whether there was/is a problem with the old …

(Copied to the list at Dan's request.)




Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?93a02a57-9f7b-473a-a63f-1c541fdd2225>