Yup that is a standard log size...
I am running with that now.
I have 5 source tree's on one xfs file system all running
make clean depend bzImage modules
Plus I have your doio test running another file system, sucking down
all my memory I might add.
6:11pm up 4 days, 45 min, 5 users, load average: 8.64, 6.57, 4.82
81 processes: 74 sleeping, 7 running, 0 zombie, 0 stopped
CPU states: 41.5% user, 58.4% system, 0.0% nice, 0.0% idle
Mem: 126116K av, 124060K used, 2056K free, 0K shrd, 256K buff
Swap: 136512K av, 24748K used, 111764K free 75112K cached
PID USER PRI NI SIZE RSS SHARE STAT LIB %CPU %MEM TIME COMMAND
22553 cattelan 15 0 167M 167M 167M R 0 18.4 136.1 69:22 doio
22554 cattelan 15 0 165M 165M 165M R 0 18.4 134.2 69:23 doio
27382 cattelan 2 0 444 444 304 R 0 3.7 0.3 0:00 mkdep
27385 cattelan 2 0 472 472 332 R 0 3.4 0.3 0:00 mkdep
27392 cattelan 3 0 1456 1456 464 S 0 3.4 1.1 0:00 make
27387 cattelan 1 0 448 448 308 D 0 3.0 0.3 0:00 mkdep
27409 cattelan 9 0 1280 1280 464 S 0 2.0 1.0 0:00 make
3 root 2 0 0 0 0 SW 0 1.6 0.0 4:04 kflushd
2 root 0 0 0 0 0 DW 0 1.5 0.0 1:14 kswapd
27415 cattelan 19 0 1276 1276 440 R 0 1.5 1.0 0:00 make
So far I haven't seen any pb_hold counts < 2 messages.
At Tue, 25 Apr 2000 15:20:23 -0700,
Rajagopal Ananthanarayanan <ananth@xxxxxxx> wrote:
>
>
>
> As suggested in the meeting:
>
> ---------
> /sbin/mkfs_xfs -f -b size=4096 /dev/sdb1
> meta-data=/dev/sdb1 isize=256 agcount=8, agsize=131030 blks
> data = bsize=4096 blocks=1048233, imaxpct=25
> = sunit=0 swidth=0 blks, unwritten=1
> naming =version 2 bsize=4096
> log =internal log bsize=4096 blocks=1200
> realtime =none extsz=65536 blocks=0, rtextents=0
> arch =1 mode=NATIVE, SIM MODE=NATIVE
> --------------------
>
>
> --------------------------------------------------------------------------
> Rajagopal Ananthanarayanan ("ananth")
> Member Technical Staff, SGI.
> --------------------------------------------------------------------------
|