Hello guys, that is my first post to this list.
We are using aufs in a squid + dansguardian box, and thing go really
slow around 4:00pm. We already dissmissed bandwidth bottlenecks, and we
have a lot of RAM in this box (around 4GB).
The ASYNC IO counters looks very different from what we supposed to see:
ASYNC IO Counters:
Operation # Requests
open 16466
close 1
cancel 16466
write 0
read 20392
stat 0
unlink 1128
check_callback 635129
queue 0
Threads Status:
# ID # Requests
1 0x4247fbb0 2378
2 0x4227ebb0 2381
3 0x4207dbb0 2361
4 0x41e7cbb0 2358
5 0x41c7bbb0 2384
6 0x41a7abb0 2367
7 0x41879bb0 2352
8 0x41678bb0 2363
9 0x41477bb0 2380
10 0x41276bb0 2367
11 0x41075bb0 2371
12 0x40e74bb0 2365
13 0x40c73bb0 2398
14 0x40a72bb0 2394
15 0x40871bb0 2379
16 0x40670bb0 2389
Shoudn't the close requests be equal to the cancel requests?
TIA
Received on Wed May 17 2006 - 13:35:56 MDT
This archive was generated by hypermail pre-2.1.9 : Thu Jun 01 2006 - 12:00:02 MDT