Subject: High interrupt CPU usage with amanda backup on a SPARCclassic
To: None <port-sparc@netbsd.org>
From: Francis Devereux <francis@devrx.org>
List: port-sparc
Date: 02/06/2003 21:55:36
Hi,
I have a SPARCclassic that I am using as an amanda backup server. It has an
external 1GB SCSI disk and HP C1537A DDS-3 tape drive. It is running NetBSD
1.6.
When the amanda is backing up over the network the CPU usage is very high
(see top output below), mainly interrupt and system, and the tape keeps
stopping and starting. Is this normal, or is there any way I can make the
backup go faster?
Thanks,
Francis
load averages: 2.27, 2.35, 2.32
21:38:14
27 processes: 2 runnable, 24 sleeping, 1 on processor
CPU states: 1.0% user, 0.0% nice, 23.5% system, 72.5% interrupt, 2.9% idle
Memory: 16M Act, 6584K Inact, 156K Wired, 2176K Exec, 17M File, 884K Free
Swap: 128M Total, 2544K Used, 126M Free
PID USERNAME PRI NICE SIZE RES STATE TIME WCPU CPU COMMAND
627 backup 2 4 424K 924K netio 26:52 26.37% 26.37% dumper
631 backup 70 4 148K 1824K RUN 11:09 21.68% 21.68% taper
626 backup 35 0 120K 1988K RUN 4:50 6.40% 6.40% taper
2522 francis 29 0 204K 924K CPU 0:00 1.64% 0.98% top
2514 francis 10 0 636K 1276K wait 0:01 0.88% 0.88% bash
2435 backup 2 0 580K 376K netio 1:55 0.83% 0.83% gzip
5 root 18 0 0K 14M syncer 0:29 0.05% 0.05% [ioflush]
191 root 2 0 388K 732K select 1:22 0.00% 0.00% sshd
123 root 10 0 0K 14M nfsidl 0:07 0.00% 0.00% [nfsio]
4 root -18 0 0K 14M reaper 0:06 0.00% 0.00% [reaper]
3 root -18 0 0K 14M pgdaemon 0:03 0.00% 0.00%
[pagedaemon]
124 root 10 0 0K 14M nfsidl 0:02 0.00% 0.00% [nfsio]
6 root -18 0 0K 14M aiodoned 0:01 0.00% 0.00% [aiodoned]
126 root 10 0 0K 14M nfsidl 0:00 0.00% 0.00% [nfsio]
125 root 10 0 0K 14M nfsidl 0:00 0.00% 0.00% [nfsio]
205 root 10 0 248K 400K nanoslee 0:00 0.00% 0.00% cron
1 root 10 0 412K 136K wait 0:00 0.00% 0.00% init