[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[Xen-users] Re: Bad performance with SATA disk



Hi there
I'm having exactly the same problem on the same hardware. I used iozone to compare disk performance figures, and found that native (i.e. Fedora Core 3 kernel 2.6.10-1.770_FC3smp) throughput of the SATA drive outperforms xen throughput by a factor of roughly 3. Logs and config info follow below, I'll gladly attach anything else that may be of use.
Help will be greatly appreciated, thanks a bunch.
Andres

lsmod (recompiled dom0 kernel with closely as possible non-xen kernel config options):
Module                  Size  Used by
iptable_filter          2304  0
ip_tables              19968  1 iptable_filter
intel_mch_agp           8208  0
agpgart                28072  1 intel_mch_agp
ata_piix                6916  3
libata                 40452  1 ata_piix
aic7xxx               212056  0
sd_mod                 12688  5
scsi_mod               76616  3 libata,aic7xxx,sd_mod

hdparm /dev/sdb:
/dev/sdb:
IO_support   =  0 (default 16-bit)
readonly     =  0 (off)
readahead    = 256 (on)
geometry     = 19452/255/63, sectors = 160000000000, start = 0
Same response under non-xen kernel.

hdparm -d 1 /dev/sdb:
/dev/sdb:
setting using_dma to 1 (on)
HDIO_SET_DMA failed: Inappropriate ioctl for device
However, I get the same response under non-xen kernel.

My lspci and dmesg output are very similar to those of Martti,
lspci | grep SATA:
00:1f.2 IDE interface: Intel Corp. 6300ESB SATA Storage Controller (rev 02)

dmesg:
ata1: SATA max UDMA/133 cmd 0x170 ctl 0x376 bmdma 0xFEA8 irq 15
ata1: dev 0 cfg 49:2f00 82:346b 83:7f01 84:4003 85:3469 86:3e01 87:4003 88:207f
ata1: dev 0 ATA, max UDMA/133, 78125000 sectors: lba48
ata1: dev 1 cfg 49:2f00 82:3469 83:7f61 84:4003 85:3469 86:3e41 87:4003 88:207f
ata1: dev 1 ATA, max UDMA/133, 312500000 sectors: lba48
ata1: dev 0 configured for UDMA/133
ata1: dev 1 configured for UDMA/133
scsi2 : ata_piix
 Vendor: ATA       Model: ST340014AS        Rev: 8.05
 Type:   Direct-Access                      ANSI SCSI revision: 05
SCSI device sda: 78125000 512-byte hdwr sectors (40000 MB)
SCSI device sda: drive cache: write back
SCSI device sda: 78125000 512-byte hdwr sectors (40000 MB)
SCSI device sda: drive cache: write back
sda: sda1 sda2
Attached scsi disk sda at scsi2, channel 0, id 0, lun 0
 Vendor: ATA       Model: WDC WD1600JD-75H  Rev: 08.0
 Type:   Direct-Access                      ANSI SCSI revision: 05
SCSI device sdb: 312500000 512-byte hdwr sectors (160000 MB)
SCSI device sdb: drive cache: write back
SCSI device sdb: 312500000 512-byte hdwr sectors (160000 MB)
SCSI device sdb: drive cache: write back
sdb: sdb1
Attached scsi disk sdb at scsi2, channel 0, id 1, lun 0


Hi!


I have few Dell PowerEdge 750 servers (each with 4 GB RAM and 250 GB SATA disk)
in our lab. The disk controllers and the disks are identified like this:


# lspci | grep SATA
0000:00:1f.2 IDE interface: Intel Corp. 6300ESB SATA Storage Controller (rev 02)


ata1: SATA max UDMA/133 cmd 0x170 ctl 0x376 bmdma 0xFEA8 irq 15
ata1: dev 0 cfg 49:2f00 82:3469 83:7f61 84:4003 85:3469 86:3e41 87:4003 88:207f
ata1: dev 0 ATA, max UDMA/133, 488281250 sectors: lba48
ata1: dev 0 configured for UDMA/133
scsi0 : ata_piix
 Vendor: ATA       Model: WDC WD2500JD-75H  Rev: 08.0
 Type:   Direct-Access                      ANSI SCSI revision: 05
SCSI device sda: 488281250 512-byte hdwr sectors (250000 MB)
SCSI device sda: drive cache: write back
SCSI device sda: 488281250 512-byte hdwr sectors (250000 MB)
SCSI device sda: drive cache: write back
sda: sda1 sda2 sda3
Attached scsi disk sda at scsi0, channel 0, id 0, lun 0


When running bonnie with the normal Linux 2.6.11 kernel I get numbers like
these:


Version 1.03 ------Sequential Output------ --Sequential Input- --Random-
                   -Per Chr- --Block-- -Rewrite- -Per Chr- --Block-- --Seeks--
Machine        Size K/sec %CP K/sec %CP K/sec %CP K/sec %CP K/sec %CP  /sec %CP
xen2             8G 29093  98 56445   6 20761   3 25706  88 53444   3 176.3   0
                   ------Sequential Create------ --------Random Create--------
                   -Create-- --Read--- -Delete-- -Create-- --Read--- -Delete--
             files  /sec %CP  /sec %CP  /sec %CP  /sec %CP  /sec %CP  /sec %CP
                16  4081  99 +++++ +++ +++++ +++  4125  99 +++++ +++ 12994  99


So far so good.


If I now I reboot to the Xen domain-0 (Xen 2.0.6 and Linux 2.6.11) and run
bonnie again (without any active domain-U) I get these numbers:


Version 1.03 ------Sequential Output------ --Sequential Input- --Random-
                   -Per Chr- --Block-- -Rewrite- -Per Chr- --Block-- --Seeks--
Machine        Size K/sec %CP K/sec %CP K/sec %CP K/sec %CP K/sec %CP  /sec %CP
xen2             2G  5798  95 50965  53 23156  15  6450  98 55912  16 128.1   1
                   ------Sequential Create------ --------Random Create--------
                   -Create-- --Read--- -Delete-- -Create-- --Read--- -Delete--
             files  /sec %CP  /sec %CP  /sec %CP  /sec %CP  /sec %CP  /sec %CP
                16   664  99 +++++ +++ +++++ +++   676  99 +++++ +++  2699  99


Now, the block output and input seem to be almost same. However, "Per Chr" and
sequential and random file creation are very slow compared to the normal Linux
kernel. The disk should still be in UDMA/133 mode with domain-0:


ata1: SATA max UDMA/133 cmd 0x170 ctl 0x376 bmdma 0xFEA8 irq 15
ata1: dev 0 cfg 49:2f00 82:3469 83:7f61 84:4003 85:3469 86:3e41 87:4003 88:207f
ata1: dev 0 ATA, max UDMA/133, 488281250 sectors: lba48
ata1: dev 0 configured for UDMA/133
scsi0 : ata_piix
 Vendor: ATA       Model: WDC WD2500JD-75H  Rev: 08.0
 Type:   Direct-Access                      ANSI SCSI revision: 05


Any ideas why this difference is so huge with domain-0?


Martti


_______________________________________________
Xen-users mailing list
Xen-users@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-users


 


Rackspace

Lists.xenproject.org is hosted with RackSpace, monitoring our
servers 24x7x365 and backed by RackSpace's Fanatical Support®.