Date: Wed, 8 Nov 2000 16:35:13 -0800 (PST) From: Brian Behlendorf <brian@collab.net> To: stable@freebsd.org Subject: disk I/O Message-ID: <Pine.BSF.4.21.0011081625011.1844-100000@yez.hyperreal.org>
next in thread | raw e-mail | index | archive | help
RELENG_4, kernel last updated & built Aug 31st, so apologies if this
touches a problem that was fixed since then. I'd be happy to upgrade
after 4.2-RELEASE goes out if people think there was a fix applied
for this.
I'm having agonizingly slow disk I/O. Granted, this is a heavily pounded
machine (~2M web hits/day with ~50K being CGI or java, 500K mail
deliveries using qmail, lots of CVS accesses) yet I get spikes in the load
going up to 200 at times, during which the machine's largely frozen to
shell access & other services are much slower. It always comes back (go
freebsd!) but it still feels like I/O is the culprit. The CPU almost
always has idle time available, yet during the day even doing an "ls" in a
moderately sized dir is slow. Here's iostat 5:
tty da0 da1 acd0 cpu
tin tout KB/t tps MB/s KB/t tps MB/s KB/t tps MB/s us ni sy in id
2 59 0.00 0 0.00 0.00 0 0.00 0.00 0 0.00 23 3 14 3 55
0 171 4.45 91 0.40 9.62 118 1.11 0.00 0 0.00 65 0 27 7 0
0 633 5.81 110 0.62 10.38 142 1.44 0.00 0 0.00 50 0 24 4 21
0 435 4.00 85 0.33 6.90 157 1.06 0.00 0 0.00 73 0 18 5 4
0 635 4.25 91 0.38 8.80 101 0.87 0.00 0 0.00 74 0 22 4 0
0 495 4.36 92 0.39 7.67 156 1.17 0.00 0 0.00 60 0 37 3 0
0 450 4.39 95 0.41 8.58 147 1.23 0.00 0 0.00 56 0 27 4 13
0 254 5.15 90 0.45 7.38 170 1.23 0.00 0 0.00 5 0 8 3 83
0 530 4.21 83 0.34 7.99 171 1.33 0.00 0 0.00 8 0 15 3 74
0 502 4.13 93 0.37 7.17 169 1.19 0.00 0 0.00 21 0 16 5 58
relevant dmesg output:
sym0: <895> port 0x9000-0x90ff mem 0xfa201000-0xfa201fff,0xfa202000-0xfa2020ff irq 10 at device 4.0 on pci1
sym0: Symbios NVRAM, ID 7, Fast-40, LVD, parity checking
sym0: open drain IRQ line driver, using on-chip SRAM
sym0: using LOAD/STORE-based firmware.
da1 at sym0 bus 0 target 1 lun 0
da1: <IBM DMVS18D 0077> Fixed Direct Access SCSI-3 device
da1: 80.000MB/s transfers (40.000MHz, offset 31, 16bit), Tagged Queueing Enabled
da1: 17501MB (35843670 512 byte sectors: 255H 63S/T 2231C)
da0 at sym0 bus 0 target 0 lun 0
da0: <HP 18.2GB A 80-B001 B001> Fixed Direct Access SCSI-2 device
da0: 80.000MB/s transfers (40.000MHz, offset 15, 16bit), Tagged Queueing Enabled
da0: 17366MB (35566480 512 byte sectors: 255H 63S/T 2213C)
Ideas? This is actually during a more lightly loaded time period. It
just "feels" like I should be able to push more than 80-100 tps through
da0 (lots of small writes and fsyncs, admittedly) and more than 150tps on
da1 (which is almost all reads). Anything else I should be looking at?
Brian
To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-stable" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?Pine.BSF.4.21.0011081625011.1844-100000>