Archive

Posts Tagged ‘storage’

I/O performance and stripe_cache_size, contd.

June 8, 2010 1 comment

There are some anomalies in the previous graphs, notably the incongruent results for stripe cache sizes of 768 and 32768. So I ran some more tests with a smoother range of  I/O blocksize.

NB : Each variable was tested three times, with all caches synced and flushed before and after. The average of those three tests were used to plot the data points on the graphs.

Caveat : I also reduced the test file size to 650MB (I know, I know, very bad practice to change multiple variables whilst testing).

I/O performance and stripe_cache_size

June 6, 2010 2 comments

Results from a basic synthetic benchmark of I/O performance whilst varying the stripe_cache_size tunable under the sysfs tree at /sys/block/md0/md/stripe_cache_size

Tests were performed on a QNAP SS-439 NAS  :

  • Intel Atom N270 (1.6GHz) 82801G
  • 2GB RAM
  • Linux 2.6.30.6 kernel
  • CFQ I/O elevator
  • RAID5  (128kb chunksize) – 4* WD10TPVT 1TB drives (4kb physical sectors, aka. Advanced Format)
  • EXT3 filesystem (noatime)

Whilst reading and writing in blocks of just 512 bytes, there seems to be no discernible benefit in setting a larger stripe cache size, with read performance dropping marginally as the cache size is increased.

The first interesting results appeared when the blocksize was increased to 4096bytes. Read performance drops off sharply as we increase the cache, though write performance gains a small amount.

At a blocksize of 1MB, our previous findings are reinforced. Read performance decreases significantly once past very low cache sizes, though write performance benefits a small amount by using larger values for the stripe_cache_size.

WD Passport Essential SE 1TB – Advanced format?

June 2, 2010 3 comments

Having been using this WD Passport Essential SE 1TB USB removable drive (formatted as ext3) to backup my NAS for the last few months, I’ve been consistently underwhelmed by it’s performance. So I tried aligning the sectors to a 4kb boundary on the off chance it behaved like the WD10TPVT drives in use within the NAS itself, and I’m seeing write performance increases of such magnitude I’m left thinking the drive within (a WD10TMVV model) has 4kb physical sectors as well, Western Digitals so called ‘Advanced Format’.

A very simple benchmark of I/O write performance with non-4kb-aligned-sectors:

# fdisk -lu /dev/sds

Disk /dev/sds: 999.5 GB, 999501594624 bytes
255 heads, 63 sectors/track, 121515 cylinders, total 1952151552 sectors
Units = sectors of 1 * 512 = 512 bytes

   Device Boot      Start         End      Blocks   Id  System
/dev/sds1   *          63  1952138474   976069206   83  Linux
# time dd if=/dev/zero of=zero count=1000 bs=1M
1000+0 records in
1000+0 records out
1048576000 bytes (1.0 GB) copied, 65.5702 s, 16.0 MB/s

real    1m5.869s
user    0m0.015s
sys     0m8.167s

And again with 4kb-aligned-sectors:

# fdisk -lu /dev/sds

Disk /dev/sds: 999.5 GB, 999501594624 bytes
255 heads, 63 sectors/track, 121515 cylinders, total 1952151552 sectors
Units = sectors of 1 * 512 = 512 bytes

   Device Boot      Start         End      Blocks   Id  System
/dev/sds1   *           8  1952151551   976075772   83  Linux
# time dd if=/dev/zero of=./zero count=1000 bs=1M
1000+0 records in
1000+0 records out
1048576000 bytes (1.0 GB) copied, 36.2553 s, 28.9 MB/s

real    0m36.577s
user    0m0.009s
sys     0m7.906s

A near 100% improvement in write performance!

WesternDigital WD10TPVT – Advanced Format

April 3, 2010 4 comments

These are details from the WD10TPVT, a 2.5” 1TB (12.5mm) SATA drive from Western Digital.

It’s the Advanced Format (4096-byte physical sectors) version of their WD10TEVT disk, though with it’s current firmware you might not know it except for the small warning label on the top sticker of the drive itself.

This label suggests every OS other than Windows XP will work with this drive without any special adjustments, but I’d like to know if this is really true, or if there is a tangible performance impact of using this drive with a default partition layout on other operating systems, notably Linux.

Tejun Heo (of LKML fame) posted a very informative wiki article about the 4kb sector issue here : https://ata.wiki.kernel.org/index.php/ATA_4_KiB_sector_issues

SMART details :

Device Model:     WDC  WD10TPVT-00HT5T0
Serial Number:    W  -DXW151CU0xxxx
Firmware Version: 01.01A01
User Capacity:     1,000,204,886,016 bytes

Linux detection :

scsi 6:0:0:0:  Direct-Access     WDC      WD10TPVT-00HT5T0 01.0 PQ: 0 ANSI: 3
sd 6:0:0:0: [sdd]  1953525168 512-byte hardware sectors: (1.00  TB/931 GiB)
sd 6:0:0:0: [sdd] Write  Protect is off
sd 6:0:0:0: [sdd] Mode  Sense: 23 00 10 00
sd 6:0:0:0: [sdd] Write  cache: enabled, read cache: enabled, supports DPO and FUA

Oddly, WD seem to have purposefully made the drive advertise itself as having only 512-byte sectors, surely this is going to cause confusion with software trying to correctly identify and partition these 4096-byte sectored Advanced Format disks?

WesternDigital Passport Essential SE 1TB

November 24, 2009 Leave a comment
The drive contained within is a WD10TMVV, SATA bridged to USB 2.5” 1TB disk, a portable version of the currently elusive WD10TEVT, though interestingly this drive has a native USB port rather than the more familiar SATA connection.

It comes formatted as NTFS :

Disk /dev/sds: 999.5 GB,  999501594624 bytes
255 heads, 63  sectors/track, 121515 cylinders
Units = cylinders of 16065 *  512 = 8225280 bytes
Device Boot       Start         End      Blocks   Id  System
/dev/sds1                1      121516   976074752    7  HPFS/NTFS

SMART details :

Device Model:     WDC  WD10TMVV-11A27S0
Serial Number:     WD-WX30A990xxxx
Firmware Version: 01.01A01
User Capacity:     1,000,204,886,016 bytes

Linux detection :

scsi 10:0:0:0:  Direct-Access     WD       My Passport 070A 1030 PQ: 0 ANSI: 4
sd 10:0:0:0: [sds]  1952151552 512-byte hardware sectors: (999 GB/930 GiB)
sd 10:0:0:0: [sds] Write  Protect is off
sd 10:0:0:0: [sds] Mode  Sense: 23 00 10 00
sd 10:0:0:0: [sds] Assuming  drive cache: write through
sd 10:0:0:0: [sds] Assuming  drive cache: write through
sds:<5>sd 10:0:0:0:  Attached scsi generic sg18 type 0
scsi 10:0:0:1:  CD-ROM            WD       Virtual CD 070A  1030 PQ: 0 ANSI: 4
sd 10:0:0:0: [sds] Attached  SCSI disk
scsi 10:0:0:1: Attached  scsi generic sg19 type 5
scsi 10:0:0:2:  Enclosure         WD       SES Device       1030 PQ: 0 ANSI: 4
scsi 10:0:0:2: Attached  scsi generic sg20 type 13

Hmm, there seems to be a large discrepancy between the SMART size details and the partition table, probably the ‘Virtual CD’ above that WD seem to have cobbled into the drive firmware ?

On WDC’s support page I find this which claims to disable the ‘Virtual CD’ crapware, perhaps useful for anyone frustrated by having a CD appear each time they plug in the drive :

Mac : http://www.wdc.com/wdproducts/updates/?family=wdsmartwareutilitiesmac
Windows : http://www.wdc.com/wdproducts/updates/?family=wdsmartwareutilities

Indeed, after running the above tools, the ‘Virtual CD’ is now hidden from view :

scsi 11:0:0:0:  Direct-Access     WD       My Passport 070A 1032 PQ: 0 ANSI: 4
sd 11:0:0:0: [sds]  1952151552 512-byte hardware sectors: (999 GB/930 GiB)
sd 11:0:0:0: [sds] Write  Protect is off
sd 11:0:0:0: [sds] Mode  Sense: 23 00 10 00
sd 11:0:0:0: [sds] Assuming  drive cache: write through
sd 11:0:0:0: [sds] Assuming  drive cache: write through
sds:<5>sd 11:0:0:0:  Attached scsi generic sg18 type 0
sds1
scsi 11:0:0:1:  Enclosure         WD       SES Device       1032 PQ: 0 ANSI: 4
sd 11:0:0:0: [sds] Attached  SCSI disk
scsi 11:0:0:1: Attached  scsi generic sg19 type 13

Though the capacity remains the same with approximately 700MB lost to unwanted WesternDigital crapware, so this 1TB drive is in fact not really 1TB at all.

Tags: ,

NAS considerations

December 20, 2008 1 comment
I’d been searching for my ideal home NAS system, here were a few of my candidates.

In the end I plumped for the QNAP SS-439 because of it’s low power requirements and I had a few 2.5” disks to spare anyway. I voided the warranty almost immediately by opening it up, see here.

QNAP SS-439 Pro Turbo NAS Datasheet Website

++ Intel Atom N270 1.6GHz (3200 bogomips)
++ QPKG packages can be installed
++ Dual 1000Mbit
++ Native iSCSI support
++ Upgradable RAM (2GB) (see here for internal photos)
— QNAP are known violators of the GPL by not releasing kernel source code
— Atom CPU not fast enough for >18MB/s via SSH/SCP/SFTP.
— Atom CPU not fast enough for >50MB/s via FTP/SMB/AFP.
— 2.5” SATA drives only come in <=500GB sizes unless they’re 12.5mm.
— It’s not as expensive as the rest, with no drives it’s about £500UKP)
18W in use, at 14p/KWH for 24h*365days = £22.08 per annum
15W sleep, at 14p/KWH for 24h*365days = £18.40 per annum

QNAP TS-439 Pro Turbo NAS Datasheet Website

++ Intel Atom based system
++ QPKG packages can be installed
++ Dual 1000Mbit
++ Native iSCSI support
— QNAP are known violators of the GPL by not releasing kernel source code
— Atom CPU not fast enough for >18MB/s via SSH/SCP/SFTP.
— Atom CPU not fast enough for >50MB/s via FTP/SMB/AFP.
— Non-upgradable RAM (1GB)
— It’s even more expensive, again no drives (>£600UKP)
39W in use, at 14p/KWH for 24h*365days = £50.10 per annum
29W sleep, at 14p/KWH for 24h*365days = £37.25 per annum

Netgear ReadyNAS NV+ Datasheet Website

++ 1000Mbit
— Slow 32bit SPARC system
— It’s power consumption is too high.
— It’s expensive, even without any drives (>£500UKP)
55W typical, at 14p/KWH for 24h*365days = £70.66 per annum

Netgear ReadyNAS Duo Datasheet Website

++ Upgradable RAM (up to 1GB)
++ 1000Mbit
— High power consumption
— No Wake On LAN
— No native iSCSI support
— Slow 32bit SPARC system (IT3107 @ 186 bogomips)
35W in use, at 14p/KWH for 24h*365days = £42.92 per annum
27W typical, at 14p/KWH for 24h*365days = £33.11 per annum
12W sleep, at 14p/KWH for 24h*365days = £14.71 per annum

Cobalt Qube 3 Pro Plus Datasheet Website

++ I have two
++ Strongbolt2 ongoing development
— 512MB max. (possible to use 1GB, but >512MB uncachable)
— Slow AMD K6-III 450MHz CPU
— Only two PATA (max 120GB) disks
— 100Mbit
45W max, at 14p/KWH for 24h*365days = £57.81 per annum
33W typical, at 14p/KWH for 24h*365days = £42.39 per annum
22W idle, at 14p/KWH for 24h*365days = £28.26 per annum

DIY NAS
Atom 230 / 330 CPU (4W / 8W) (SingleCore / DualCore)
Core 2 Duo SU9300 (10W)
++ Possibly cheaper
++ Opensolaris
++ ZFS
— Difficult to ensure low power consumption
— No Atom PCI-E motherboards?

Tags:

Upgrade MacBook Pro to 200GB 7200rpm

September 14, 2007 Leave a comment

Having now several Parallels Virtual Machine images I’d quickly reached the limits of the 5400rpm 120GB SATA hard disk my MacBook Pro was CTO’d with.

To replace it I chose the the Seagate Momentus 7200.2 200GB model with 16MB cache.

Replacing the drive is time-consuming but straight forward, I could have used CarbonCopyCloner, but plumped for SuperDuper instead to copy the contents of the old drive to the new one which I had placed temporarily in a SATA Firewire enclosure.

The new disk isn’t really much faster, but the extra space made the operation worthwhile.