Date: Mon, 16 Nov 1998 16:06:09 +0300 (MSK) From: "Cyril A. Vechera" <cyril@main.piter.net> To: freebsd-questions@FreeBSD.ORG Subject: strange disk I/O overload when free memory is large enough. how to fixc? Message-ID: <199811161306.QAA17527@main.piter.net>
next in thread | raw e-mail | index | archive | help
Hi,
I've met a problem, FreeBSD 2.2.6-RELEASE.
On this system i run network database server wich operates with
hash(3) and btree(3) tables, and perl5 client with C-lib via XSUB.
in this case dbopen was done with NULL file (table in memory)
The problem is that when i run massive records insertion,
the database server is slows down and sometimes ps shows that
db-server is marked with 'D':
USER PID %CPU %MEM VSZ RSS TT STAT STARTED TIME COMMAND
netdbm 626 11.8 1.2 520 768 p3 D+ 3:31PM 0:14.58 netdbm
and after half of second
netdbm 626 11.8 1.2 520 768 p3 I+ 3:31PM 0:14.80 netdbm
vmstat says when client is not running
procs memory page faults cpu
r b w avm fre flt re pi po fr sr w0 in sy cs us sy id
0 0 0 5340 35144 1 0 0 0 0 0 0 232 17 1 0 1 99
and when client works
procs memory page faults cpu
r b w avm fre flt re pi po fr sr w0 in sy cs us sy id
0 1 0 8544 34592 207 0 0 0 26 0 74 1432 5324 445 21 24 55
iostat reports about 2000 sps on wd0 (this machine has only wd0)
there are no any NFS or same stuff.
netdbm doesn't use any page protection mechanizm, but how can you see
there is enough free memory
vmstat -s says that
0 swap pager pageins
0 swap pager pages paged in
0 swap pager pageouts
0 swap pager pages paged out
348 vnode pager pageins
1591 vnode pager pages paged in
0 vnode pager pageouts
0 vnode pager pages paged out
what is the problem?
Sincerely your,
Cyril A. Vechera
email:cyril@piter.net --------- http://sply.piter.net
To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-questions" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199811161306.QAA17527>