home *** CD-ROM | disk | FTP | other *** search
- Xref: sparky comp.sys.sun.admin:6240 comp.unix.admin:4980
- Newsgroups: comp.sys.sun.admin,comp.unix.admin
- Path: sparky!uunet!mcsun!sun4nl!eur.nl!pk
- From: pk@cs.few.eur.nl (Paul Kranenburg)
- Subject: Re: biff not biffing
- Message-ID: <1992Sep9.123044.27574@cs.few.eur.nl>
- Sender: news@cs.few.eur.nl
- Reply-To: pk@cs.few.eur.nl
- Organization: Erasmus University Rotterdam
- References: <1992Aug29.190229.19212@europa.asd.contel.com> <1992Aug29.202158.22016@europa.asd.contel.com> <1992Aug29.211845.21326@ccu.umanitoba.ca> <BtxGD1.1sM@cs.psu.edu>
- Date: Wed, 9 Sep 1992 12:30:44 GMT
- Lines: 32
-
- In <BtxGD1.1sM@cs.psu.edu> kenh@leps5.phys.psu.edu (Ken Hornstein) writes:
-
- >In article <1992Aug29.211845.21326@ccu.umanitoba.ca> mills@ccu.umanitoba.ca (Gary Mills) writes:
- >>In <1992Aug29.202158.22016@europa.asd.contel.com> pascoe@rocky.gte.com (Dave Pascoe) writes:
- >>
- >>>commented out comsat in /etc/inetd.conf and sent a kill -1 to inetd
- >>>and then uncommented comsat and sent another kill -1.
- >>
- >>>The comsat daemon was out to lunch for some reason. This is a common
- >>>problem on machines which handle a lot of mail, since the comsat
- >>>daemon can tend to get confused at to what state it's in when a SIGHUP
- >>>to inetd comes along.
- >>
- >>I have seen this same behavior with other daemons started from inetd,
- >>and I suspect that the bug is in inetd. It apparently assumes that
- >>a particular daemon is running when it rereads its config file and then
- >>refuses to start another one.
-
- >I've seen this happen before with tftpd. At our site we use tftpd to download
- >fonts to our X terminals. More than once I was adding or deleting a service
- >to inetd.conf and when I sent inetd the HUP signal it refused to run tftpd
- >again ("Hey, why do I suddenly get the error message 'Cannot load 8x12'?").
- >Whoops :-) Killing and restarting inetd fixed the problem.
-
- It definitely is a bug in Sun's inetd. It doesn't honour the "wait" field
- when processing a HUP signal (while the service is "active").
-
- I have reported this bug a long time ago to our local Sun helpdesk (Netherlands)
- but never received a fix. I ended up modifying the BSD inetd to replace Sun's
- (we needed other functionality anyway).
-
- -pk
-