Hello
May be that I have a jfs or kernel problem:
I oserve the following behavior.
Systemload grows very fast and a few seconds later the system doesn't
react any longer.
The server is a big Samba Server with 900 GByte jfs filesystem
configured on a raid 5 disk.
So I have some trouble with my users.
See the enries in the /var/log/messages.
The kernel version:
frasdat39:/tmp # uname -a
Linux frasdat39 2.4.9 #9 SMP Sat Oct 20 19:46:20 CEST 2001 i686 unknown
JFS Version 1.0.3
I tested it with jfs.1.0.7 too.
In heavy load situations jfscommit mutate to a Zombie prozess. Because
this is a child of init, you can only power down the system.
I am using a Software raid5 with 6 178GByte Disks
frasdat39:/service/disk2/spethus # df
Filesystem 1k-blocks Used Available Use% Mounted on
/dev/sda3 7763888 3348008 4021488 46% /
/dev/sda1 23333 6606 15523 30% /boot
/dev/md1 886405896 450466380 435939516 51%
/service/raiddisk1
frasdat39:/service/disk2/spethus # cat /proc/mdstat
Personalities : [raid5]
read_ahead 1024 sectors
md1 : active raid5 sdq1[6] sdp1[5] sdo1[4] sdn1[3] sdm1[2] sdl1[1]
sdk1[0]
886465920 blocks level 5, 32k chunk, algorithm 2 [6/6] [UUUUUU]
unused devices: <none>
_______________________________________________
Jfs-discussion mailing list
[EMAIL PROTECTED]
http://www-124.ibm.com/developerworks/oss/mailman/listinfo/jfs-discussion
