dear ILUG user group,

i have trouble to crank up the write speed on a 4-disc RAID-0 system i
installed recently.  i use two Promise IDE controller cards in
addition to the motherboard IDE controller.  writing to the /dev/md0
(linux software RAID-0) appears to be also somewhat strange not just
slow: writing of a big file or of many small files occurs in a burst
like mode, usually at 72mB/s every 1-2 seconds for very short time,
1/2 second or less, observed visually with xosview and ultimately it
takes double the time to write to the /dev/md0 as compared to any of
the other regular discs.  interestingly, when applying many write
processes (through multiple matlab processes. e.g.) the 'burst rate'
simply goes up, meaning the average write speed seems to increase
proportionally(?).  an additional observation of importance is that
during such a 'burst-mode' /dev/md0 write attempt the 4 logical
processors (SMP) go to sleep, are just dead for a second or longer.

but, reading from the /dev/md0 is fast, no problem.  reading/writing
to the first 3 regular HDs is very fast too.

i defined all IRQs & IOs at boot time, within GRUB's menu.lst file.
kernel is a 2.4.20 tailored to my needs, on Slackware8.1, all
filesystems of the reiserfs type.  all 80-pol UDMA cabling used.  at
boot /sbin/hdparm gets applied with options like '-d1' and '-S 241' to
make all discs UDMA and keep them cool.  raidtools are from the slack
installation, with raidstart v0.3d compiled for md raidtools-0.90.
kernel config parameters are:

        CONFIG_MD=y
        CONFIG_BLK_DEV_MD=m
        CONFIG_MD_RAID0=m
        CONFIG_PDC202XX_BURST=y
        CONFIG_PDC202XX_FORCE=y

i am aware that i use two different makes of HDs of identical size
within the RAID-0 system but i just didn't have the money/time to deal
with that aspect.

any comments or advice will be appreciated much.


                        pisti



here my hardware setup:

        motherboard SuperMicro P4DCE+ with 2 IDE channels:

                /dev/hda        HD
                /dev/hda        CD-R
                /dev/hdc        DVD-RW
                /dev/hdd        CD_RW

        Promise Ultra100 Tx2 with 2 channels, IRQ=18:

                /dev/hde        HD
                /dev/hdg        HD

        Promise FastTrack100 Tx4 with 4 channels, IRQ=17:

                /dev/hdi        HD
                /dev/hdk        HD
                /dev/hdm        HD
                /dev/hdo        HD


some selected boot messages:

        Uniform Multi-Platform E-IDE driver Revision: 6.31
        ide: Assuming 33MHz system bus speed for PIO modes; override with idebus=xx
        ICH2: IDE controller on PCI bus 00 dev f9
        ICH2: detected chipset, but driver not compiled in!
        ICH2: chipset revision 4
        ICH2: not 100% native mode: will probe irqs later
            ide0: BM-DMA at 0xf000-0xf007, BIOS settings: hda:DMA, hdb:DMA
            ide1: BM-DMA at 0xf008-0xf00f, BIOS settings: hdc:DMA, hdd:DMA
        PDC20268: IDE controller on PCI bus 03 dev 08
        PDC20268: chipset revision 2
        PDC20268: not 100% native mode: will probe irqs later
        PDC20268: (U)DMA Burst Bit ENABLED Primary MASTER Mode Secondary MASTER Mode.
            ide2: BM-DMA at 0xb000-0xb007, BIOS settings: hde:pio, hdf:pio
            ide3: BM-DMA at 0xb008-0xb00f, BIOS settings: hdg:pio, hdh:pio
        PDC20270: IDE controller on PCI bus 05 dev 08
        PDC20270: chipset revision 2
        PDC20270: not 100% native mode: will probe irqs later
        PDC20270: (U)DMA Burst Bit ENABLED Primary MASTER Mode Secondary MASTER Mode.
            ide4: BM-DMA at 0x7000-0x7007, BIOS settings: hdi:pio, hdj:pio
            ide5: BM-DMA at 0x7008-0x700f, BIOS settings: hdk:pio, hdl:pio
        PDC20270: IDE controller on PCI bus 05 dev 10
        PDC20270: chipset revision 2
        PDC20270: not 100% native mode: will probe irqs later
        PDC20270: ROM enabled at 0x000dc000
        PDC20270: (U)DMA Burst Bit ENABLED Primary MASTER Mode Secondary MASTER Mode.
            ide6: BM-DMA at 0x8400-0x8407, BIOS settings: hdm:pio, hdn:pio
            ide7: BM-DMA at 0x8408-0x840f, BIOS settings: hdo:pio, hdp:pio

        hda: ST380021A, ATA DISK drive
        hdb: LG CD-ROM CRD-8522B, ATAPI CD/DVD-ROM drive
        hdc: PIONEER DVD-RW DVR-105, ATAPI CD/DVD-ROM drive
        hdd: RICOH CD-R/RW MP7083A, ATAPI CD/DVD-ROM drive
        hde: ST380021A, ATA DISK drive
        hdg: WDC WD400BB-00AUA1, ATA DISK drive
        hdi: ST380021A, ATA DISK drive
        hdk: WDC WD800JB-00CRA1, ATA DISK drive
        hdm: WDC WD800JB-00CRA1, ATA DISK drive
        hdo: WDC WD800JB-00CRA1, ATA DISK drive

        ide0 at 0x1f0-0x1f7,0x3f6 on irq 14
        ide1 at 0x170-0x177,0x376 on irq 15
        ide2 at 0xa000-0xa007,0xa402 on irq 18
        ide3 at 0xa800-0xa807,0xac02 on irq 18
        ide4 at 0x6000-0x6007,0x6402 on irq 17
        ide5 at 0x6800-0x6807,0x6c02 on irq 17
        ide6 at 0x7400-0x7407,0x7802 on irq 17
        ide7 at 0x7c00-0x7c07,0x8002 on irq 17

        hda: 156301488 sectors (80026 MB) w/2048KiB Cache, CHS=9729/255/63
        blk: queue c02dfb8c, I/O limit 4095Mb (mask 0xffffffff)
        hde: 156301488 sectors (80026 MB) w/2048KiB Cache, CHS=155061/16/63, UDMA(33)
        blk: queue c02dfef0, I/O limit 4095Mb (mask 0xffffffff)
        hdg: 78165360 sectors (40021 MB) w/2048KiB Cache, CHS=77545/16/63, UDMA(33)
        blk: queue c02e0254, I/O limit 4095Mb (mask 0xffffffff)
        hdi: 156301488 sectors (80026 MB) w/2048KiB Cache, CHS=155061/16/63, UDMA(33)
        blk: queue c02e05b8, I/O limit 4095Mb (mask 0xffffffff)
        hdk: 156301488 sectors (80026 MB) w/8192KiB Cache, CHS=155061/16/63, UDMA(33)
        blk: queue c02e091c, I/O limit 4095Mb (mask 0xffffffff)
        hdm: 156301488 sectors (80026 MB) w/8192KiB Cache, CHS=155061/16/63, UDMA(33)
        blk: queue c02e0c80, I/O limit 4095Mb (mask 0xffffffff)
        hdo: 156301488 sectors (80026 MB) w/8192KiB Cache, CHS=155061/16/63, UDMA(33)

        Partition check:
          hda: hda1 hda2 hda3 hda4 < hda5 hda6 >
          hde: [PTBL] [9729/255/63] hde1 hde2 hde3 hde4 < hde5 hde6 >
          hdg: [PTBL] [4865/255/63] hdg1 hdg2 hdg3
          hdi: unknown partition table
          hdk: unknown partition table
          hdm: unknown partition table
          hdo: unknown partition table

        md: md driver 0.90.0 MAX_MD_DEVS=256, MD_SB_DISKS=27
        md: raid0 personality registered as nr 2

        ide0: BM-DMA at 0xf000-0xf007, BIOS settings: hda:DMA, hdb:DMA
        ide1: BM-DMA at 0xf008-0xf00f, BIOS settings: hdc:DMA, hdd:DMA
        ide2: BM-DMA at 0xb000-0xb007, BIOS settings: hde:pio, hdf:pio
        ide3: BM-DMA at 0xb008-0xb00f, BIOS settings: hdg:pio, hdh:pio
        ide4: BM-DMA at 0x7000-0x7007, BIOS settings: hdi:pio, hdj:pio
        ide5: BM-DMA at 0x7008-0x700f, BIOS settings: hdk:pio, hdl:pio
        ide6: BM-DMA at 0x8400-0x8407, BIOS settings: hdm:pio, hdn:pio
        ide7: BM-DMA at 0x8408-0x840f, BIOS settings: hdo:pio, hdp:pio


after doing a '/sbin/raidstart' & 'mount /dev/md0':

        blk: queue c02df4c4, I/O limit 4095Mb (mask 0xffffffff)
        blk: queue c02dfb8c, I/O limit 4095Mb (mask 0xffffffff)
        blk: queue c02dfef0, I/O limit 4095Mb (mask 0xffffffff)
        blk: queue c02e0254, I/O limit 4095Mb (mask 0xffffffff)
        blk: queue c02e05b8, I/O limit 4095Mb (mask 0xffffffff)
        blk: queue c02e091c, I/O limit 4095Mb (mask 0xffffffff)
        blk: queue c02e0c80, I/O limit 4095Mb (mask 0xffffffff)
         [events: 0000003c]
         [events: 0000003c]
         [events: 0000003c]
         [events: 0000003c]
        md: autorun ...
        md: considering hdo ...
        md:  adding hdo ...
        md:  adding hdm ...
        md:  adding hdk ...
        md:  adding hdi ...
        md: created md0
        md: bind<hdi,1>
        md: bind<hdk,2>
        md: bind<hdm,3>
        md: bind<hdo,4>
        md: running: <hdo><hdm><hdk><hdi>
        md: hdo's event counter: 0000003c
        md: hdm's event counter: 0000003c
        md: hdk's event counter: 0000003c
        md: hdi's event counter: 0000003c
        md0: max total readahead window set to 992k
        md0: 4 data-disks, max readahead per data-disk: 248k
        raid0: looking at hdi
        raid0:   comparing hdi(78150656) with hdi(78150656)
        raid0:   END
        raid0:   ==> UNIQUE
        raid0: 1 zones
        raid0: looking at hdk
        raid0:   comparing hdk(78150656) with hdi(78150656)
        raid0:   EQUAL
        raid0: looking at hdm
        raid0:   comparing hdm(78150656) with hdi(78150656)
        raid0:   EQUAL
        raid0: looking at hdo
        raid0:   comparing hdo(78150656) with hdi(78150656)
        raid0:   EQUAL
        raid0: FINAL 1 zones
        raid0: zone 0
        raid0: checking hdi ... contained as device 0
          (78150656) is smallest!.
        raid0: checking hdk ... contained as device 1
        raid0: checking hdm ... contained as device 2
        raid0: checking hdo ... contained as device 3
        raid0: zone->nb_dev: 4, size: 312602624
        raid0: current zone offset: 78150656
        raid0: done.
        raid0 : md_size is 312602624 blocks.
        raid0 : conf->smallest->size is 312602624 blocks.
        raid0 : nb_zone is 1.
        raid0 : Allocating 8 bytes for hash.
        md: updating md0 RAID superblock on device
        md: hdo [events: 0000003d]<6>(write) hdo's sb offset: 78150656
        md: hdm [events: 0000003d]<6>(write) hdm's sb offset: 78150656
        md: hdk [events: 0000003d]<6>(write) hdk's sb offset: 78150656
        md: hdi [events: 0000003d]<6>(write) hdi's sb offset: 78150656
        md: ... autorun DONE.
        reiserfs: checking transaction log (device 09:00) ...
        Using r5 hash to sort names
        ReiserFS version 3.6.25


---
[EMAIL PROTECTED]


=================================================================
To unsubscribe, send mail to [EMAIL PROTECTED] with
the word "unsubscribe" in the message body, e.g., run the command
echo unsubscribe | mail [EMAIL PROTECTED]

Reply via email to