Date: Wed, 13 Dec 2017 16:11:20 +0100 From: "Hartmann, O." <o.hartmann@walstatt.org> To: Cy Schubert <Cy.Schubert@komquats.com> Cc: "O. Hartmann" <ohartmann@walstatt.org>, "Rodney W. Grimes" <freebsd-rwg@pdx.rh.CN85.dnsmgr.net>, FreeBSD CURRENT <freebsd-current@freebsd.org>, Freddie Cash <fjwcash@gmail.com>, Alan Somers <asomers@freebsd.org> Subject: Re: SMART: disk problems on RAIDZ1 pool: (ada6:ahcich6:0:0:0): CAMstatus: ATA Status Error Message-ID: <20171213161116.1889f178@hermann> In-Reply-To: <20171212225827.CBCF2266@spqr.komquats.com> References: <20171212225827.CBCF2266@spqr.komquats.com>
next in thread | previous in thread | raw e-mail | index | archive | help
On Tue, 12 Dec 2017 14:58:28 -0800
Cy Schubert <Cy.Schubert@komquats.com> wrote:
> There are a couple of ways you can address this. You'll need to
> offline the vdev first. If you've done a smartcrl -t long and if the
> test failed, smartcrl -a will tell you which block it had an issue
> with. You can use dd, ddrescue or dd_rescue to dd the block over
> itself. The drive may rewrite the (weak) block or if it fails to it
> will remap it (subsequently showing as reallocated).
>
> Of course there is a risk. If the sector is any of the boot blocks
> there is a good chance the server will hang.
The drive is part of a dedicated storage-only pool. The boot drive is a
fast SSD. So I do not care about this - well, to say it more politely:
I do not have to take care of that aspect.
>
> You have to be *absolutely* sure which the bad sector is. And, there
> may be more. There is a risk of data loss.
>
> I've used this technique many times. Most times it works perfectly.
> Other times the affected file is lost but the rest of the file system
> is recovered. And again there is always the risk.
>
> Replace the disk immediately if you experience a growing succession
> of pending sectors. Otherwise replace the disk at your earliest
> convenience.
The ZFS scrubbing of the volume ended this morning, leaving the pool in
a healthy state. After reboot, there was no sign of CAM errors again.
But there is something else I'm worried about. The mainboard I use is a
ASRock Z77 Pro4-M.
The board has a cripple Intel MCP with 6 SATA ports from the chipset,
two of them SATA 6GB, 4 SATA II, and one additional chip with two SATA
6GB ports:
[...]
ahci0@pci0:2:0:0: class=0x010601 card=0x06121849 chip=0x06121b21
rev=0x01 hdr=0x00 vendor = 'ASMedia Technology Inc.'
device = 'ASM1062 Serial ATA Controller'
class = mass storage
subclass = SATA
bar [10] = type I/O Port, range 32, base 0xe050, size 8, enabled
bar [14] = type I/O Port, range 32, base 0xe040, size 4, enabled
bar [18] = type I/O Port, range 32, base 0xe030, size 8, enabled
bar [1c] = type I/O Port, range 32, base 0xe020, size 4, enabled
bar [20] = type I/O Port, range 32, base 0xe000, size 32, enabled
bar [24] = type Memory, range 32, base 0xf7b00000, size 512,
enabled
[...]
Attached to that ASM1062 SATA chip, is a backup drive via eSATA
connector, a WD 4 TB RED drive. It seems, whenever I attach this drive
and it is online, I experience problems on the ZFS pool, which is
attached to the MCP SATA ports.
Is this possible? I mean, as I asked before, a weird/defect cabling
would trigger different error schemes (CRC errors). Due to the fact
that the external drive is physically decoupled and is not capable of
coupling in vibrations, bad sector errors seem to me unlikely. But this
is simply a though of someone without special knowledge about physics
of HDDs.
I think people responding to my thread made it clear that the WD Green
isn't the first-choice-solution for a 20/6 (not 24/7) duty drive and
the fact, that they have serviced now more than 25000 hours, it would
be wise to replace them with alternatives.
>
> If using a zfs mirror (not in your case) detatch and attach will
> rewrite any weakly written sectors and reallocate pending sectors.
>
> ---
> Sent using a tiny phone keyboard.
> Apologies for any typos and autocorrect.
> Also, this old phone only supports top post. Apologies.
>
> Cy Schubert
> <Cy.Schubert@cschubert.com> or <cy@freebsd.org>
> The need of the many outweighs the greed of the few.
> ---
>
> -----Original Message-----
> From: O. Hartmann
> Sent: 12/12/2017 14:19
> To: Rodney W. Grimes
> Cc: O. Hartmann; FreeBSD CURRENT; Freddie Cash; Alan Somers
> Subject: Re: SMART: disk problems on RAIDZ1 pool:
> (ada6:ahcich6:0:0:0): CAMstatus: ATA Status Error
>
> Am Tue, 12 Dec 2017 10:52:27 -0800 (PST)
> "Rodney W. Grimes" <freebsd-rwg@pdx.rh.CN85.dnsmgr.net> schrieb:
>
>
> Thank you for answering that fast!
>
> > > Hello,
> > >
> > > running CURRENT (recent r326769), I realised that smartmond sends
> > > out some console messages when booting the box:
> > >
> > > [...]
> > > Dec 12 14:14:33 <3.2> box1 smartd[68426]: Device: /dev/ada6, 1
> > > Currently unreadable (pending) sectors Dec 12 14:14:33 <3.2> box1
> > > smartd[68426]: Device: /dev/ada6, 1 Offline uncorrectable sectors
> > > [...]
> > >
> > > Checking the drive's SMART log with smartctl (it is one of four
> > > 3TB disk drives), I gather these informations:
> > >
> > > [... smartctl -x /dev/ada6 ...]
> > > Error 42 [17] occurred at disk power-on lifetime: 25335 hours
> > > (1055 days + 15 hours) When the command that caused the error
> > > occurred, the device was active or idle.
> > >
> > > After command completion occurred, registers were:
> > > ER -- ST COUNT LBA_48 LH LM LL DV DC
> > > -- -- -- == -- == == == -- -- -- -- --
> > > 40 -- 51 00 00 00 00 c2 7a 72 98 40 00 Error: UNC at LBA =
> > > 0xc27a7298 = 3262804632
> > >
> > > Commands leading to the command that caused the error were:
> > > CR FEATR COUNT LBA_48 LH LM LL DV DC Powered_Up_Time
> > > Command/Feature_Name -- == -- == -- == == == -- -- -- -- --
> > > --------------- -------------------- 60 00 b0 00 88 00 00 c2 7a
> > > 73 20 40 08 23:38:12.195 READ FPDMA QUEUED 60 00 b0 00 80 00
> > > 00 c2 7a 72 70 40 08 23:38:12.195 READ FPDMA QUEUED 2f 00 00
> > > 00 01 00 00 00 00 00 10 40 08 23:38:12.195 READ LOG EXT 60
> > > 00 b0 00 70 00 00 c2 7a 73 20 40 08 23:38:09.343 READ FPDMA
> > > QUEUED 60 00 b0 00 68 00 00 c2 7a 72 70 40 08 23:38:09.343
> > > READ FPDMA QUEUED [...]
> > >
> > > and
> > >
> > > [...]
> > > SMART Attributes Data Structure revision number: 16
> > > Vendor Specific SMART Attributes with Thresholds:
> > > ID# ATTRIBUTE_NAME FLAGS VALUE WORST THRESH FAIL
> > > RAW_VALUE 1 Raw_Read_Error_Rate POSR-K 200 200 051
> > > - 64 3 Spin_Up_Time POS--K 178 170 021
> > > - 6075 4 Start_Stop_Count -O--CK 098 098 000
> > > - 2406 5 Reallocated_Sector_Ct PO--CK 200 200 140
> > > - 0 7 Seek_Error_Rate -OSR-K 200 200 000 -
> > > 0 9 Power_On_Hours -O--CK 066 066 000 - 25339
> > > 10 Spin_Retry_Count -O--CK 100 100 000 - 0
> > > 11 Calibration_Retry_Count -O--CK 100 100 000 - 0
> > > 12 Power_Cycle_Count -O--CK 098 098 000 - 2404
> > > 192 Power-Off_Retract_Count -O--CK 200 200 000 - 154
> > > 193 Load_Cycle_Count -O--CK 001 001 000 -
> > > 2055746 194 Temperature_Celsius -O---K 122 109 000
> > > - 28 196 Reallocated_Event_Count -O--CK 200 200 000
> > > - 0 197 Current_Pending_Sector -O--CK 200 200 000
> > > - 1 198 Offline_Uncorrectable ----CK 200 200 000
> > > - 1 199 UDMA_CRC_Error_Count -O--CK 200 200 000
> > > - 0 200 Multi_Zone_Error_Rate ---R-- 200 200 000
> > > - 5 ||||||_ K auto-keep
> > > |||||__ C event count
> > > ||||___ R error rate
> > > |||____ S speed/performance
> > > ||_____ O updated online
> > > |______ P prefailure warning
> > >
> > > [...]
> >
> > The data up to this point informs us that you have 1 bad sector
> > on a 3TB drive, that is actually an expected event given the data
> > error rate on this stuff is such that your gona have these now
> > and again.
> >
> > Given you have 1 single event I would not suspect that this drive
> > is dying, but it would be prudent to prepare for that possibility.
>
> Hello.
>
> Well, I copied simply "one single event" that has been logged so far.
>
> As you (and I) can see, it is error #42. After I posted here, a
> reboot has taken place because the "repair" process on the Pool
> suddenly increased time and now I'm with error #47, but
> interestingly, it is a new block that is damaged, but the SMART
> attribute fields show this for now:
>
> [...]
> SMART Attributes Data Structure revision number: 16
> Vendor Specific SMART Attributes with Thresholds:
> ID# ATTRIBUTE_NAME FLAGS VALUE WORST THRESH FAIL RAW_VALUE
> 1 Raw_Read_Error_Rate POSR-K 200 200 051 - 69
> 3 Spin_Up_Time POS--K 178 170 021 - 6075
> 4 Start_Stop_Count -O--CK 098 098 000 - 2406
> 5 Reallocated_Sector_Ct PO--CK 200 200 140 - 0
> 7 Seek_Error_Rate -OSR-K 200 200 000 - 0
> 9 Power_On_Hours -O--CK 066 066 000 - 25343
> 10 Spin_Retry_Count -O--CK 100 100 000 - 0
> 11 Calibration_Retry_Count -O--CK 100 100 000 - 0
> 12 Power_Cycle_Count -O--CK 098 098 000 - 2404
> 192 Power-Off_Retract_Count -O--CK 200 200 000 - 154
> 193 Load_Cycle_Count -O--CK 001 001 000 - 2055746
> 194 Temperature_Celsius -O---K 122 109 000 - 28
> 196 Reallocated_Event_Count -O--CK 200 200 000 - 0
> 197 Current_Pending_Sector -O--CK 200 200 000 - 0
> 198 Offline_Uncorrectable ----CK 200 200 000 - 1
> 199 UDMA_CRC_Error_Count -O--CK 200 200 000 - 0
> 200 Multi_Zone_Error_Rate ---R-- 200 200 000 - 5
> ||||||_ K auto-keep
> |||||__ C event count
> ||||___ R error rate
> |||____ S speed/performance
> ||_____ O updated online
> |______ P prefailure warning
> [...]
>
>
> 197 Current_Pending_Sector decreased to zero so far, but with every
> reboot, the error count seems to increase:
>
>
> [...]
> Error 47 [22] occurred at disk power-on lifetime: 25343 hours (1055
> days + 23 hours) When the command that caused the error occurred, the
> device was active or idle.
>
> After command completion occurred, registers were:
> ER -- ST COUNT LBA_48 LH LM LL DV DC
> -- -- -- == -- == == == -- -- -- -- --
> 40 -- 51 00 00 00 00 c2 19 d9 88 40 00 Error: UNC at LBA =
> 0xc219d988 = 3256473992
>
> Commands leading to the command that caused the error were:
> CR FEATR COUNT LBA_48 LH LM LL DV DC Powered_Up_Time
> Command/Feature_Name -- == -- == -- == == == -- -- -- -- --
> --------------- -------------------- 60 00 b0 00 d0 00 00 c2 19 da
> 28 40 08 1d+07:12:34.336 READ FPDMA QUEUED 60 00 b0 00 c8 00 00 c2
> 19 d9 78 40 08 1d+07:12:34.336 READ FPDMA QUEUED 2f 00 00 00 01 00
> 00 00 00 00 10 40 08 1d+07:12:34.336 READ LOG EXT 60 00 b0 00 b8 00
> 00 c2 19 da 28 40 08 1d+07:12:31.484 READ FPDMA QUEUED 60 00 b0 00
> b0 00 00 c2 19 d9 78 40 08 1d+07:12:31.483 READ FPDMA QUEUED
>
>
> I think this is watching a HDD dying, isn't it?
>
> I'd say, a broken cabling would produce different errors, wouldn't it?
>
> The Western Digital Green series HDD is a useful fellow when the HDD
> is used as a single drive. I think there might be an issue with
> paring 4 HDDs, 3 of them "GREEN", in a RAIDZ and physically sitting
> next to each other. Maybe it is time to replace them one by one ...
>
> >
> >
> > >
> > > The ZFS pool is RAIDZ1, comprised of 3 WD Green 3TB HDD and one
> > > WD RED 3 TB HDD. The failure occured is on one of the WD Green 3
> > > TB HDD.
> > Ok, so the data is redundantly protected. This helps a lot.
> >
> > > The pool is marked as "resilvered" - I do scrubbing on a regular
> > > basis and the "resilvering" message has now aapeared the second
> > > time in row. Searching the net recommend on SMART attribute 197
> > > errors, in my case it is one, and in combination with the
> > > problems occured that I should replace the disk.
> >
> > It is probably putting the RAIDZ in that state as the scrub is
> > finding a block it can not read.
> >
> > >
> > > Well, here comes the problem. The box is comprised from
> > > "electronical waste" made by ASRock - it is a Socket
> > > 1150/IvyBridge board, which has its last Firmware/BIOS update got
> > > in 2013 and since then UEFI booting FreeBSD from a HDD isn't
> > > possible (just to indicate that I'm aware of having issues with
> > > crap, but that is some other issue right now). The board's SATA
> > > connectors are all populated.
> > >
> > > So: Due to the lack of adequate backup space I can only
> > > selectively backup portions, most of the space is occupied by
> > > scientific modelling data, which I had worked on. So backup
> > > exists! In one way or the other. My concern is how to replace the
> > > faulty HDD! Most HowTo's indicate a replacement disk being
> > > prepared and then "replaced" via ZFS's replace command. This
> > > isn't applicable here.
> > >
> > > Question: is it possible to simply pull the faulty disk (implies
> > > I know exactly which one to pull!) and then prepare and add the
> > > replacement HDD and let the system do its job resilvering the
> > > pool?
> >
> > That may work, but I think I have a simpler solution.
> >
> > >
> > > Next question is: I'm about to replace the 3 TB HDD with a more
> > > recent and modern 4 TB HDD (WD RED 4TB). I'm aware of the fact
> > > that I can only use 3 TB as the other disks are 3 TB, but I'd
> > > like to know whether FreeBSD's ZFS is capable of handling it?
> >
> > Someone else?
> >
> > >
> > > This is the first time I have issues with ZFS and a faulty drive,
> > > so if some of my questions sound naive, please forgive me.
> >
> > One thing to try is to see if we can get the drive to fix itself,
> > first order of business is can you take this server out of
> > service? If so I would simply try to do a
> > repeat 100 dd if=/dev/whicheverhdisbad of=/dev/null conv=noerror,
> > sync iseek=3262804632
> >
> > That is trying to read that block 100 times, if it successful even
> > 1 time smart should remap the block and you are all done.
>
> Given the fact, that this errorneous block is like a moving target,
> it this solution still the favorite one? I'll try, but I already have
> the replacement 4 TB HDD at hand.
>
> >
> > If that fails we can try to zero the block, there is a risk here,
> > but raidz should just handle this as a data corruption of a block.
> > This could possibly lead to data loss, so USE AT YOUR OWN RISK
> > ASSESMENT. dd if=/dev/zero of=/dev/whateverdrivehasissues bs=512
> > count=1 oseek=3262804632
>
> I would then be oseek=3256473992, too.
>
> >
> > That should forceable overwrite the bad block with 0's, the smart
> > firmware well see this in the pending list, write the data, read it
> > back, if successful remove it from the pending list, if failed
> > reallocate the block and write the 0's to the reallocation and add
> > 1 to the remapped block count.
> >
> > You might google for "how to fix a pending reallocation"
> >
> > > Thanks in advance,
> > > Oliver
> > > --
> > > O. Hartmann
> >
>
> Kind regards,
>
> Oliver
>
>
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20171213161116.1889f178>