High load during drive read/write : dmesg output, need help understanding what they are

[Keith Richie]

I'm having issues with HD reading and writing using too high of a commit
level. Meaning when I write large amounts of data to a hard drive, the
system slows to a crawl. If I only had one HD, this would be understandable,
but I'm using 3 HDs. If I attempt to write a large file to /dev/hde
(storage), and then launch a program which resides on /dev/hda, the system
halts to a crawl.

Copy and move from /dev/hdb to dev/hde have the same performance problems.
I've tried formatting hdb and hde as JFS, XFS, EXT2/3 and Reiser. They have
the same issues, with JFS being the best case, but not acceptable.

DMA is enabled on all drives
AMD 3000+
NForce3 using PATA, no SATA drives
1gb RAM
Ubuntu 6.06 w/K7 kernel.

hdparm -tT /dev/hda

/dev/hdb:
 Timing cached reads:   2560 MB in  2.00 seconds = 1278.92 MB/sec
 Timing buffered disk reads:  164 MB in  3.02 seconds =  54.22 MB/sec

hdparm -tT /dev/hdb

/dev/hdb:
 Timing cached reads:   2560 MB in  2.00 seconds = 1278.28 MB/sec
 Timing buffered disk reads:  166 MB in  3.01 seconds =  55.18 MB/sec


 hdparm -tT /dev/hde

/dev/hde:
 Timing cached reads:   2552 MB in  2.00 seconds = 1276.19 MB/sec
 Timing buffered disk reads:  164 MB in  3.02 seconds =  54.24 MB/sec



I found some interesting parts in dmesg, not sure what they mean though.

[4296067.739000] Mem-info:
[4296067.739000] DMA per-cpu:
[4296067.739000] cpu 0 hot: low 0, high 0, batch 1 used:0
[4296067.739000] cpu 0 cold: low 0, high 0, batch 1 used:0
[4296067.739000] DMA32 per-cpu: empty
[4296067.739000] Normal per-cpu:
[4296067.739000] cpu 0 hot: low 0, high 186, batch 31 used:15
[4296067.739000] cpu 0 cold: low 0, high 62, batch 15 used:56
[4296067.739000] HighMem per-cpu:
[4296067.739000] cpu 0 hot: low 0, high 42, batch 7 used:2
[4296067.739000] cpu 0 cold: low 0, high 14, batch 3 used:12
[4296067.739000] Free pages:        5304kB (344kB HighMem)
[4296067.739000] Active:137837 inactive:100583 dirty:6506 writeback:6075
unstable:0 free:1326 slab:9626 mapped:135557 pagetables:724
[4296067.739000] DMA free:3548kB min:68kB low:32kB high:68kB active:2924kB
inactive:4848kB present:16384kB pages_scanned:33 all_unreclaimable? no
[4296067.739000] lowmem_reserve[]: 0 0 880 1007
[4296067.739000] DMA32 free:0kB min:0kB low:0kB high:0kB active:0kB
inactive:0kB present:0kB pages_scanned:0 all_unreclaimable? no
[4296067.739000] lowmem_reserve[]: 0 0 880 1007
[4296067.739000] Normal free:1412kB min:3756kB low:1876kB high:3756kB
active:483124kB inactive:341920kB present:901120kB pages_scanned:1108
all_unreclaimable? no
[4296067.739000] lowmem_reserve[]: 0 0 0 1023
[4296067.739000] HighMem free:344kB min:128kB low:168kB high:336kB
active:65300kB inactive:55564kB present:131008kB pages_scanned:0
all_unreclaimable? no
[4296067.739000] lowmem_reserve[]: 0 0 0 0
[4296067.739000] DMA: 13*4kB 1*8kB 0*16kB 1*32kB 0*64kB 1*128kB 1*256kB
0*512kB 1*1024kB 1*2048kB 0*4096kB = 3548kB
[4296067.739000] DMA32: empty
[4296067.739000] Normal: 41*4kB 0*8kB 0*16kB 1*32kB 1*64kB 1*128kB 0*256kB
0*512kB 1*1024kB 0*2048kB 0*4096kB = 1412kB
[4296067.739000] HighMem: 52*4kB 1*8kB 0*16kB 2*32kB 1*64kB 0*128kB 0*256kB
0*512kB 0*1024kB 0*2048kB 0*4096kB = 344kB
[4296067.739000] Swap cache: add 4611, delete 4611, find 2/4, race 0+0
[4296067.739000] Free swap  = 1033828kB
[4296067.739000] Total swap = 1052216kB
[4296067.739000] Free swap:       1033828kB
[4296067.739000] 262128 pages of RAM
[4296067.739000] 32752 pages of HIGHMEM
[4296067.739000] 3584 reserved pages
[4296067.739000] 68670 pages shared
[4296067.739000] 0 pages swap cached
[4296067.739000] 6506 pages dirty
[4296067.739000] 6075 pages writeback
[4296067.739000] 135557 pages mapped
[4296067.739000] 9626 pages slab
[4296067.739000] 724 pages pagetables
[4296067.739000] eth0: too many iterations (6) in nv_nic_irq.
[4296067.774000] swapper: page allocation failure. order:0, mode:0x20
[4296067.774000]  [<c01524d9>] __alloc_pages+0x229/0x330
[4296067.774000]  [<c0155f3b>] kmem_getpages+0x4b/0xe0
[4296067.774000]  [<c0156f68>] cache_grow+0xb8/0x200
[4296067.774000]  [<c01572c8>] cache_alloc_refill+0x218/0x260
[4296067.774000]  [<c015763e>] __kmalloc+0x7e/0x80
[4296067.774000]  [<c028c1ff>] __alloc_skb+0x5f/0x180
[4296067.774000]  [<f89207b2>] nv_alloc_rx+0x102/0x130 [forcedeth]
[4296067.774000]  [<f8922141>] nv_nic_irq+0xf1/0x1a0 [forcedeth]
[4296067.774000]  [<f89221f0>] nv_do_nic_poll+0x0/0x70 [forcedeth]
[4296067.774000]  [<f892223d>] nv_do_nic_poll+0x4d/0x70 [forcedeth]
[4296067.774000]  [<c012f2f5>] run_timer_softirq+0xd5/0x1d0
[4296067.774000]  [<c012a5b6>] __do_softirq+0x76/0xe0
[4296067.774000]  [<c012a655>] do_softirq+0x35/0x40
[4296067.774000]  [<c012a735>] irq_exit+0x45/0x50
[4296067.774000]  [<c0105c0e>] do_IRQ+0x1e/0x30
[4296067.774000]  [<c0103eb6>] common_interrupt+0x1a/0x20
[4296067.774000]  [<c0101030>] default_idle+0x0/0x60
[4296067.774000]  [<c010105c>] default_idle+0x2c/0x60
[4296067.774000]  [<c010111b>] cpu_idle+0x6b/0xb0
[4296067.774000]  [<c03a89fa>] start_kernel+0x19a/0x200
[4296067.774000]  [<c03a8370>] unknown_bootoption+0x0/0x1e0
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.ubuntu.com/archives/ubuntu-users/attachments/20060616/585c2455/attachment.html>