4k sector drives
Quoted 5 lines
I have these:
Model=WDC WD20EARS-00MVWB0, FwRev=50.0AB50, SerialNo=WD-WMAZ20035646 Model=WDC WD20EARS-00MVWB0, FwRev=50.0AB50, SerialNo=WD-WMAZ20028252Any chance to get Linux know they are 4k sector drives manually?
Your revision numbers are different than mine, because you have the 2TB drive, but i guess its the same for these too. You can try "hdparm -I /dev/sdX" (capital i) or "cat /sys/block/sdX/queue/hw_sector_size" to see what they report.
Mine report "Logical/Physical Sector size: 512 bytes"
There are newer revisions that report "Physical Sector size: 4096 bytes"
Nevertheless, you don't have to get them to report 4k to linux. As long as you align the partitions its fine. As i said in the other post, newer fdisk doesn't use "DOS compatibility mode" and uses sectors by default (instead of stupid CHS) with a 1MB alignment, so every partition will be properly aligned. If you don't dual boot Windows, you can also use GPT partitioning scheme.
Also, from what i have seen, xfsprogs are quite clever and find most stuff on their own. I usually mkfs with "-s size=4096" just to be sure.
Slightly offtopic, there is another "problem" with these drives. They park the heads every 8 seconds, so they _might_ wear prematurely. You can change the 8sec timer or disable it with the wdidle3 utility from this url
http://support.wdc.com/product/download.asp?groupid=609&sid=113&lang=en
If you want to know more about the problem, there are many forums where you can read about wdidle3 and i also posted about it on my blog.
http://lordkhelben.wordpress.com/2010/05/15/wd-caviar-green-hard-disks-idle-settings/
I hope i helped.
_______________________________________________
xfs mailing list
[email protected]
http://oss.sgi.com/mailman/listinfo/xfs
Quoted 1 line · On Mittwoch, 21. Juli 2010 Khelben Blackstaff wrote:
I hope i helped.
Sound good, will try and report back. Thx.
Signature · 13 lines
mit freundlichen Grüssen, Michael Monnerie, Ing. BSc it-management Internet Services http://proteger.at [gesprochen: Prot-e-schee] Tel: 0660 / 415 65 31 ****** Aktuelles Radiointerview! ****** http://www.it-podcast.at/aktuelle-sendung.html // Wir haben im Moment zwei Häuser zu verkaufen: // http://zmi.at/langegg/ // http://zmi.at/haus2009/
_______________________________________________
xfs mailing list
[email protected]
http://oss.sgi.com/mailman/listinfo/xfs
Quoted 8 lines · On Mittwoch, 21. Juli 2010 Khelben Blackstaff wrote:
Nevertheless, you don't have to get them to report 4k to linux. As
long as you align the partitions its fine. As i said in the other post, newer fdisk doesn't use "DOS compatibility mode" and uses sectors by default (instead of stupid CHS) with a 1MB alignment, so every partition will be properly aligned. If you don't dual boot Windows, you can also use GPT partitioning scheme. Also, from what i have seen, xfsprogs are quite clever and find most stuff on their own. I usually mkfs with "-s size=4096" just to be sure.
I always do -b size=4096 anyway, so I'm fine here. And I created the partitions starting sector 512, so that should be good also.
Quoted 7 lines
Slightly offtopic, there is another "problem" with these drives. They
park the heads every 8 seconds, so they might wear prematurely. You can change the 8sec timer or disable it with the wdidle3 utility from this url
http://support.wdc.com/product/download.asp?groupid=609&sid=113&a
mp;lang=en
Quoted 5 lines
If you want to know more about the problem, there are many forums
where you can read about wdidle3 and i also posted about it on my blog.
http://lordkhelben.wordpress.com/2010/05/15/wd-caviar-green-hard-dis
ks-idle-settings/
Thanks, I did this, and it even worked for my other WDC WD2002FYPS-01U1B0, which was on 12 seconds (the EARS were on 8s as you said), and now all three drives have a 120s delay before parking. That should be sufficient. It didn't work for the VelociRaptor in my PC, just to mention.
Signature · 13 lines
mit freundlichen Grüssen, Michael Monnerie, Ing. BSc it-management Internet Services http://proteger.at [gesprochen: Prot-e-schee] Tel: 0660 / 415 65 31 ****** Aktuelles Radiointerview! ****** http://www.it-podcast.at/aktuelle-sendung.html // Wir haben im Moment zwei Häuser zu verkaufen: // http://zmi.at/langegg/ // http://zmi.at/haus2009/
_______________________________________________
xfs mailing list
[email protected]
http://oss.sgi.com/mailman/listinfo/xfs
Quoted 10 lines · Michael Monnerie
Quoted 8 lines · On Mittwoch, 21. Juli 2010 Khelben Blackstaff wrote:
Nevertheless, you don't have to get them to report 4k to linux. As
long as you align the partitions its fine. As i said in the other post, newer fdisk doesn't use "DOS compatibility mode" and uses sectors by default (instead of stupid CHS) with a 1MB alignment, so every partition will be properly aligned. If you don't dual boot Windows, you can also use GPT partitioning scheme. Also, from what i have seen, xfsprogs are quite clever and find most stuff on their own. I usually mkfs with "-s size=4096" just to be sure.
Quoted 2 lines
I always do -b size=4096 anyway, so I'm fine here. And I created the partitions starting sector 512, so that should be good also.
-b size=4096 is the default anyway, for 4k sector drivers you also want -s size=4096, which you do not want for 512 byte sector disks.
What values do the files
/sys/block/<device>/queue/logical_block_size /sys/block/<device>/queue/physical_block_size /sys/block/<device>/alignment_offset
say about your disk?
_______________________________________________
xfs mailing list
[email protected]
http://oss.sgi.com/mailman/listinfo/xfs
Christoph Hellwig put forth on 7/23/2010 4:58 AM:
Quoted 7 lines
What values do the files
/sys/block/<device>/queue/logical_block_size /sys/block/<device>/queue/physical_block_size /sys/block/<device>/alignment_offset
say about your disk?
FWIW, for comparison, my 500GB Blue series WD5000AAKS-00V1A0 shows
logical_block_size 512physical_block_size 512alignment_offset 0
Signature · 6 lines
Stan _______________________________________________ xfs mailing list [email protected] http://oss.sgi.com/mailman/listinfo/xfs
Quoted 5 lines · Stan Hoeppner
FWIW, for comparison, my 500GB Blue series WD5000AAKS-00V1A0 shows
logical_block_size512physical_block_size512alignment_offset0
Which means there is absolutely no visible 4k sector support.
_______________________________________________
xfs mailing list
[email protected]
http://oss.sgi.com/mailman/listinfo/xfs
Christoph Hellwig put forth on 7/23/2010 5:47 AM:
Quoted 8 lines · Stan Hoeppner
Quoted 5 lines
FWIW, for comparison, my 500GB Blue series WD5000AAKS-00V1A0 shows
logical_block_size512physical_block_size512alignment_offset0Which means there is absolutely no visible 4k sector support.
Sorry for the confusion Christoph.
I meant for comparison to a plain old 512 byte/sec drive from the same manufacturer. The Western Digital Caviar Blue model WD5000AAKS, my drive above, has 512 byte sectors. His 2TB Caviar Green has 4096 byte sectors.
Signature · 6 lines
Stan _______________________________________________ xfs mailing list [email protected] http://oss.sgi.com/mailman/listinfo/xfs
Quoted 2 lines · On Freitag, 23. Juli 2010 Christoph Hellwig wrote:
-b size=4096 is the default anyway, for 4k sector drivers you also
want -s size=4096, which you do not want for 512 byte sector disks.
Thanks for clarification. Should I use "-s size=4096" despite the drive saying 512b sectors? It's a "hidden" 4K sector drive, so maybe declaring it extra for XFS helps performance? Or does it not matter at all?
Quoted 7 lines
What values do the files
/sys/block/<device>/queue/logical_block_size /sys/block/<device>/queue/physical_block_size /sys/block/<device>/alignment_offset
say about your disk?
512, 512, 0 for a Western Digital D20EARS-00MVWB0 (2TB) which has 4K sectors but obviously hides it.
Signature · 13 lines
mit freundlichen Grüssen, Michael Monnerie, Ing. BSc it-management Internet Services http://proteger.at [gesprochen: Prot-e-schee] Tel: 0660 / 415 65 31 ****** Aktuelles Radiointerview! ****** http://www.it-podcast.at/aktuelle-sendung.html // Wir haben im Moment zwei Häuser zu verkaufen: // http://zmi.at/langegg/ // http://zmi.at/haus2009/
_______________________________________________
xfs mailing list
[email protected]
http://oss.sgi.com/mailman/listinfo/xfs
Le Fri, 23 Jul 2010 12:59:07 +0200
Michael Monnerie <[email protected]> écrivait:
Quoted 4 lines
Thanks for clarification. Should I use "-s size=4096" despite the drive saying 512b sectors? It's a "hidden" 4K sector drive, so maybe declaring it extra for XFS helps performance? Or does it not matter at all?
I'd say it's the right time for a benchmark :)
Signature · 6 lines
------------------------------------------------------------------------
Emmanuel Florac | Direction technique
| Intellique
| <[email protected]>
| +33 1 78 94 84 02
------------------------------------------------------------------------_______________________________________________
xfs mailing list
[email protected]
http://oss.sgi.com/mailman/listinfo/xfs
Michael Monnerie put forth on 7/23/2010 5:59 AM:
Quoted 10 lines · On Freitag, 23. Juli 2010 Christoph Hellwig wrote:
Quoted 2 lines
-b size=4096 is the default anyway, for 4k sector drivers you also
want -s size=4096, which you do not want for 512 byte sector disks.Thanks for clarification. Should I use "-s size=4096" despite the drive saying 512b sectors? It's a "hidden" 4K sector drive, so maybe declaring it extra for XFS helps performance? Or does it not matter at all?
Quoted 2 lines
What values do the files
Quoted 8 lines
Quoted 5 lines
/sys/block/<device>/queue/logical_block_size /sys/block/<device>/queue/physical_block_size /sys/block/<device>/alignment_offset
say about your disk?
512, 512, 0 for a Western Digital D20EARS-00MVWB0 (2TB) which has 4K sectors but obviously hides it.
The WD20EARS-00MVWB0 is definitely a 4k sector drive:
http://products.wdc.com/Library/Flyer/ENG/2178-771123.pdf
If you're currently seeing somewhere between 60-120MB/s per drive with hdparm or dd then you don't need to further tweak anything. If alignment is off, from all I've read, performance will be abysmal, down in the sub 30MB/s range.
Signature · 6 lines
Stan _______________________________________________ xfs mailing list [email protected] http://oss.sgi.com/mailman/listinfo/xfs
Quoted 20 lines · Stan Hoeppner
Michael Monnerie put forth on 7/23/2010 5:59 AM:
Quoted 10 lines · On Freitag, 23. Juli 2010 Christoph Hellwig wrote:
Quoted 2 lines
-b size=4096 is the default anyway, for 4k sector drivers you also
want -s size=4096, which you do not want for 512 byte sector disks.Thanks for clarification. Should I use "-s size=4096" despite the drive saying 512b sectors? It's a "hidden" 4K sector drive, so maybe declaring it extra for XFS helps performance? Or does it not matter at all?
Quoted 2 lines
What values do the files
Quoted 8 lines
Quoted 5 lines
/sys/block/<device>/queue/logical_block_size /sys/block/<device>/queue/physical_block_size /sys/block/<device>/alignment_offset
say about your disk?
512, 512, 0 for a Western Digital D20EARS-00MVWB0 (2TB) which has 4K sectors but obviously hides it.
Quoted 6 lines
The WD20EARS-00MVWB0 is definitely a 4k sector drive:
http://products.wdc.com/Library/Flyer/ENG/2178-771123.pdfIf you're currently seeing somewhere between 60-120MB/s per drive with hdparm or dd then you don't need to further tweak anything. If alignment is off, from all I've read, performance will be abysmal, down in the sub 30MB/s range.
Only for writes, reads are for practically unaffected.
The problem for write is the "Read Modify Write" Cycle that is needed when you don't change all 8 512byte "sub-"sectors contained in a 4096 byte sector.
Bis denn
Signature · 9 lines
Real Programmers consider "what you see is what you get" to be just as bad a concept in Text Editors as it is in women. No, the Real Programmer wants a "you asked for it, you got it" text editor -- complicated, cryptic, powerful, unforgiving, dangerous. _______________________________________________ xfs mailing list [email protected] http://oss.sgi.com/mailman/listinfo/xfs
Le Fri, 23 Jul 2010 15:38:43 +0200
Matthias Schniedermeyer <[email protected]> écrivait:
Quoted 1 line
Only for writes, reads are for practically unaffected.
Sequential writes may be relatively unaffected too; if you write sequentially a couple of megabytes you'll have only a couple of additional blocks at the beginning and end of operation.
However random and small IOs should be seriously hit.
Signature · 11 lines
------------------------------------------------------------------------
Emmanuel Florac | Direction technique
| Intellique
| <[email protected]>
| +33 1 78 94 84 02
------------------------------------------------------------------------
_______________________________________________
xfs mailing list
[email protected]
http://oss.sgi.com/mailman/listinfo/xfsQuoted 8 lines · Emmanuel Florac
Le Fri, 23 Jul 2010 15:38:43 +0200
Matthias Schniedermeyer <[email protected]> écrivait:Quoted 1 line
Only for writes, reads are for practically unaffected.
Sequential writes may be relatively unaffected too; if you write sequentially a couple of megabytes you'll have only a couple of additional blocks at the beginning and end of operation.
What MAY hurt, also for large writes, are the meta-data operations while writing the file. (e.g. the file-size changes, journaling ...)
But i haven't tested that and can't say for sure how much that hurts.
Bis denn
Signature · 9 lines
Real Programmers consider "what you see is what you get" to be just as bad a concept in Text Editors as it is in women. No, the Real Programmer wants a "you asked for it, you got it" text editor -- complicated, cryptic, powerful, unforgiving, dangerous. _______________________________________________ xfs mailing list [email protected] http://oss.sgi.com/mailman/listinfo/xfs
Quoted 7 lines · Michael Monnerie
Quoted 2 lines · On Freitag, 23. Juli 2010 Christoph Hellwig wrote:
-b size=4096 is the default anyway, for 4k sector drivers you also
want -s size=4096, which you do not want for 512 byte sector disks.Thanks for clarification. Should I use "-s size=4096" despite the drive saying 512b sectors? It's a "hidden" 4K sector drive, so maybe declaring it extra for XFS helps performance? Or does it not matter at all?
If it really is one using -s size=4096 is the right thing to do.
_______________________________________________
xfs mailing list
[email protected]
http://oss.sgi.com/mailman/listinfo/xfs
Quoted 9 lines · Christoph Hellwig
Quoted 6 lines · Michael Monnerie
Quoted 2 lines · On Freitag, 23. Juli 2010 Christoph Hellwig wrote:
-b size=4096 is the default anyway, for 4k sector drivers you also
want -s size=4096, which you do not want for 512 byte sector disks.Thanks for clarification. Should I use "-s size=4096" despite the drive saying 512b sectors? It's a "hidden" 4K sector drive, so maybe declaring it extra for XFS helps performance? Or does it not matter at all?
If it really is one using -s size=4096 is the right thing to do.
Haven't read the whole thread so maybe this is redundant, but make sure all partitions (if any) are 4k aligned as well (unless it has the secret-handshake sector 63 offset...)
-Eric
_______________________________________________
xfs mailing list
[email protected]
http://oss.sgi.com/mailman/listinfo/xfs
Quoted 5 lines · On Samstag, 24. Juli 2010 Eric Sandeen wrote:
Quoted 1 line
If it really is one using -s size=4096 is the right thing to do.
Haven't read the whole thread so maybe this is redundant, but make sure all partitions (if any) are 4k aligned as well (unless it has the secret-handshake sector 63 offset...)
Thank you Christoph for clarification, also thanks to Eric - I've been aligning all partitions to a multiple of 512 sectors since a long time, which fits all stripe set sizes up to 256k. Too bad that even Linux fdisk and parted still don't do something like that, today I setup a Novell SLES 11 with XEN VM, and everywhere parted would align the first partition (when using GPT) to sector 34.. what a number.. Good that 4K sector drives arrive now, so the tools will hopefully change for a better default.
BTW: AFAIK SSD drives have a (sometimes very large) internal alignment of bytes which should be used as a "stripe size" to meet the optimal I/O size. Is there a way to find out how big that is yet?
Signature · 13 lines
mit freundlichen Grüssen, Michael Monnerie, Ing. BSc it-management Internet Services http://proteger.at [gesprochen: Prot-e-schee] Tel: 0660 / 415 65 31 ****** Aktuelles Radiointerview! ****** http://www.it-podcast.at/aktuelle-sendung.html // Wir haben im Moment zwei Häuser zu verkaufen: // http://zmi.at/langegg/ // http://zmi.at/haus2009/
_______________________________________________
xfs mailing list
[email protected]
http://oss.sgi.com/mailman/listinfo/xfs
Quoted 4 lines · Michael Monnerie
Thank you Christoph for clarification, also thanks to Eric - I've been aligning all partitions to a multiple of 512 sectors since a long time, which fits all stripe set sizes up to 256k. Too bad that even Linux fdisk and parted
Current fdisk, and with some limitations parted do the right thing and align partitions on at least a 1MB boundary. parted unfortunately lowers alignment again when any kind of optimal I/O size hint is set, but I'm still hoping for that bug to get fix in a new release.
_______________________________________________
xfs mailing list
[email protected]
http://oss.sgi.com/mailman/listinfo/xfs