From mboxrd@z Thu Jan  1 00:00:00 1970
Date: Sun,  4 Jun 2000 16:45:41 +0900
From: arisawa@ar.aichi-u.ac.jp arisawa@ar.aichi-u.ac.jp
Subject: [9fans] arpd memory leak.
Topicbox-Message-UUID: b028c982-eac8-11e9-9e20-41e7f4b1d025
Message-ID: <20000604074541.x8hwWqR90KttU2ZMpzdJcSB1TpnPCe9iZQ4gtNzbcho@z>

Hello 9fans,

In the old article, dhog@plan9.cs.su.oz.au said:
	I just noticed that the arpd on our cpu server is using 35M,
	according to ps!
	I don't have a lot of time right now to investigate.
	Does anyone know anything about this?

Now I am also in trouble!

I started cron to ping all IPs in 16 segments of our university
to get IPs of active hosts. The interval of measurement is 10 minutes.

Then,

# at the beginning:
cpu% ps
...
bootes       15    0:00   0:00    64K Read     cs
bootes       16    0:00   0:00     0K Wakeme   tcpack
bootes       17    0:00   0:00     0K Wakeme   tcpflow
bootes       25    0:00   0:00    48K Read     arpd
bootes       28    0:00   0:00    36K Read     icmp
bootes       29    0:00   0:00    36K Read     icmp

# after a hour:
cpu% ps
...
bootes       15    0:00   0:00    64K Read     cs
bootes       16    0:00   0:00     0K Wakeme   tcpack
bootes       17    0:00   0:00     0K Wakeme   tcpflow
bootes       25    0:00   0:00  6612K Read     arpd
bootes       28    0:32   0:47    48K Read     icmp
bootes       29    0:14   0:38    48K Read     icmp

# and the system hangs up in a night with the message:
	"no physical memory"

Has anyone fixed this memory leak?

Kenji Arisawa
E-mail: arisawa@aichi-u.ac.jp