(cc's added) On Mon, 18 Feb 2008 21:09:22 -0500 "David M. Strang" <dstrang@xxxxxxxxxxxxxx> wrote: > Greetings - > > A couple months back I purchased a LSI Logic MegaRAID ATA 150-4 > controller, as well as 3 Seagate 500GB SATA-II hard drives to use in my > system. Previously, I was using a pair of WD4000YR's in software raid, > which seemed to work well. I've just not gotten around to working on > migrating my data to these new drivers + controller, and it's giving me > some issues. As with most, I'm having some severe performance issues, > the performance is simply abysmal. Before getting into the details, here > is a quick overview of my configuration: > > System: > Tyan Tiger i7320/R (S5350) System Board > 2x Intel Xeon 3.0 GHz > 4GB RAM > > LSI Logic MegaRAID ATA 150-4 controller - Firmware Revision: 713S > 3x Seagate 7200.10 (Perpendicular Recording) ST3500630AS 500GB SATA-II > drives configured as a RAID-1 array with a HotSpare. > > Also, connected to the onboard controller is a WD4000YR, where all of my > data currently resides. > > I'm running Gentoo Hardended AMD64 MultiLib > (/usr/portage/profiles/hardened/amd64/multilib) > > My current kernel revision is 2.6.23-hardened-r7. > > Here are some (possibly) relevant snippets from dmesg during startup: > > ... > megaraid cmm: 2.20.2.7 (Release Date: Sun Jul 16 00:01:03 EST 2006) > megaraid: 2.20.5.1 (Release Date: Thu Nov 16 15:32:35 EST 2006) > megaraid: probe new device 0x1000:0x1960:0x1000:0x4523: bus 3:slot 3:func 0 > ACPI: PCI Interrupt 0000:03:03.0[A] -> GSI 24 (level, low) -> IRQ 24 > megaraid: fw version:[713S] bios version:[G121] > scsi0 : LSI Logic MegaRAID driver > scsi[0]: scanning scsi channel 0 [Phy 0] for non-raid devices > scsi[0]: scanning scsi channel 1 [virtual] for logical drives > scsi 0:1:0:0: Direct-Access MegaRAID LD 0 RAID1 476G 713S PQ: 0 ANSI: 2 > sd 0:1:0:0: [sda] 976762880 512-byte hardware sectors (500103 MB) > sd 0:1:0:0: [sda] Write Protect is off > sd 0:1:0:0: [sda] Mode Sense: 00 00 00 00 > sd 0:1:0:0: [sda] Asking for cache data failed > sd 0:1:0:0: [sda] Assuming drive cache: write through > sd 0:1:0:0: [sda] 976762880 512-byte hardware sectors (500103 MB) > sd 0:1:0:0: [sda] Write Protect is off > sd 0:1:0:0: [sda] Mode Sense: 00 00 00 00 > sd 0:1:0:0: [sda] Asking for cache data failed > sd 0:1:0:0: [sda] Assuming drive cache: write through > sda: sda1 sda2 sda3 sda4 > sd 0:1:0:0: [sda] Attached SCSI disk > ata_piix 0000:00:1f.2: version 2.12 > ata_piix 0000:00:1f.2: MAP [ P0 -- P1 -- ] > ACPI: PCI Interrupt 0000:00:1f.2[A] -> GSI 18 (level, low) -> IRQ 18 > PCI: Setting latency timer of device 0000:00:1f.2 to 64 > scsi1 : ata_piix > scsi2 : ata_piix > ata1: SATA max UDMA/133 cmd 0x00000000000114a0 ctl 0x000000000001149a > bmdma 0x0000000000011470 irq 18 > ata2: SATA max UDMA/133 cmd 0x0000000000011490 ctl 0x0000000000011486 > bmdma 0x0000000000011478 irq 18 > ata1.00: ATA-7: WDC WD4000YR-01PLB0, 01.06A01, max UDMA/133 > ata1.00: 781422768 sectors, multi 16: LBA48 NCQ (depth 0/32) > ata1.00: configured for UDMA/133 > scsi 1:0:0:0: Direct-Access ATA WDC WD4000YR-01P 01.0 PQ: 0 ANSI: 5 > sd 1:0:0:0: [sdb] 781422768 512-byte hardware sectors (400088 MB) > sd 1:0:0:0: [sdb] Write Protect is off > sd 1:0:0:0: [sdb] Mode Sense: 00 3a 00 00 > sd 1:0:0:0: [sdb] Write cache: enabled, read cache: enabled, doesn't > support DPO or FUA > sd 1:0:0:0: [sdb] 781422768 512-byte hardware sectors (400088 MB) > sd 1:0:0:0: [sdb] Write Protect is off > sd 1:0:0:0: [sdb] Mode Sense: 00 3a 00 00 > sd 1:0:0:0: [sdb] Write cache: enabled, read cache: enabled, doesn't > support DPO or FUA > sdb: sdb1 sdb2 sdb3 sdb4 > sd 1:0:0:0: [sdb] Attached SCSI disk > ... > > My controller is configured for Write Back Caching, Adaptive Read Ahead, > and Direct I/O (I've also tried cached I/O but it scared me...) > > The first thing I'm noticing is the horrible performance on the raid > disk, compared to the single standalone hard disk. Here is the output > from hdparm -tT on the single disk: > > -(root@server)-(~)- # hdparm -tT /dev/sdb1 > > /dev/sdb1: > Timing cached reads: 1670 MB in 2.00 seconds = 835.00 MB/sec > Timing buffered disk reads: 140 MB in 3.01 seconds = 46.45 MB/sec > > And then, the output from the raid-1 array: > > -(root@server)-(~)- # hdparm -tT /dev/sda1 > > /dev/sda1: > Timing cached reads: 1718 MB in 2.00 seconds = 859.65 MB/sec > Timing buffered disk reads: 92 MB in 3.09 seconds = 29.76 MB/sec > > I'm not sure what the deal is with the buffered disk reads being so much > WORSE than a single disk. So poor performance is a concern, but what's > more alarming are the messages showing up in DMESG. When I first tried > Cached IO - performance seemed good... except, dmesg was littered with > these errors (?): > > megaraid: aborting-14610 cmd=2a <c=1 t=0 l=0> > megaraid abort: scsi cmd:14610, do now own > megaraid: aborting-14612 cmd=2a <c=1 t=0 l=0> > megaraid abort: scsi cmd:14612, do now own > megaraid: aborting-14614 cmd=2a <c=1 t=0 l=0> > megaraid abort: scsi cmd:14614, do now own > ... > megaraid: 38 outstanding commands. Max wait 300 sec > megaraid mbox: Wait for 38 commands to complete:300 > megaraid mbox: reset sequence completed sucessfully > > I'm not certain what these mean... why am I getting aborts? > > So, I rebooted the box - and I switched back to direct I/O instead of > cached... and while not as prevelant as before, I still get the above > listed errors as well as these ones: > > megaraid abort: 14687:62[255:128], fw owner > megaraid: aborting-14689 cmd=2a <c=1 t=0 l=0> > megaraid abort: 14689:25[255:128], fw owner > megaraid: aborting-14691 cmd=2a <c=1 t=0 l=0> > megaraid abort: 14691:40[255:128], fw owner > megaraid: aborting-14693 cmd=2a <c=1 t=0 l=0> > megaraid abort: 14693:10[255:128], fw owner > megaraid: aborting-14695 cmd=2a <c=1 t=0 l=0> > megaraid abort: 14695:9[255:128], fw owner > > > I'm also a bit concerned by the dmesg output from the drive > initialization; I have it set for Write Back caching, but this shows up: > > sd 0:1:0:0: [sda] Asking for cache data failed > sd 0:1:0:0: [sda] Assuming drive cache: write through > > Why? > > I would really like to get my data over to my hardware mirror, but > frankly - I'm nervous about this controller's behavior. Other than error > messages in dmesg, and high cpu during file i/o -- it SEEMS ok, but is > it really? I somehow don't think I should be getting these types of > messages. I've searched the list archive, and I see similar messages > follow by failed resets, but my reset sequence always completes > successfully. > > Regards, > David M. Strang - To unsubscribe from this list: send the line "unsubscribe linux-scsi" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
