home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!cs.utexas.edu!swrinde!network.ucsd.edu!mvb.saic.com!tgv.com!info-multinet
- From: adelman@TGV.COM
- Newsgroups: vmsnet.networks.tcp-ip.multinet
- Subject: Re: No buffer space errors
- Message-ID: <246025DE02SEP92231714@TGV.COM>
- Date: 2 Sep 92 23:17:14 GMT
- Organization: The INFO-MULTINET Community
- Lines: 61
- X-Gateway-Source-Info: INTERNET
- X-Return-path: <info-multinet-relay@TGV.COM>
- X-RFC822-From: adelman (Kenneth Adelman) @ TGV.COM
- Nntp-Posting-Host: Mvb.Saic.Com
-
- > Yesterday, we had several strange problems that occured with one of our
- > systems running Multinet. Multinet usually runs solid as a rock, but it
- > seemed to be really hosed up and nothing we could do short of a reboot
- > could fix it.
-
- > Our system is a VAX 6420 running VMS 5.5 and we run Multinet 3.0. (We also
- > have Multinet 3.0 running under VMS 5.5 on another cluster which contains
- > a 6420 and a 6640 and both of these systems were functioning fine at the
- > time.) Yesterday we noticed that when users tried to telnet or rsh to our
- > system from other local systems in the same building, they received an error
- > message (text varies depending on where & what they were coming from)
- > indicating the system was not responding. The system was up however, it
- > hadn't been rebooted in 30+ days. When we tried to telnet from our system to
- > anywhere else, we got the message:
-
- > %MULTINET-W-ENOBUFS, No buffer space available
-
- > We also got a number of variations on the above message on our OPCOM
- > throughout the day:
- > -------------------------------------------------------------------------------
- > MultiNet Server: R_SERVICES: I/O error %MULTINET-W-ENOBUFS, No buffer space
- > available
- > MultiNet Server: R_SERVICES: Socket write error %MULTINET-W-ENOBUFS, No buffer
- > space available
- > MultiNet Server: R_SERVICES: getpeername: %MULTINET-W-ENOBUFS, No buffer space
- > available
- > MultiNet Server: Standard_Connected setsockopt KEEPALIVE: No buffer space
- > available
- > MultiNet Server: accept: No buffer space available
- > MultiNet Server: FTP: setsockopt (SO_KEEPALIVE): %MULTINET-W-ENOBUFS, No buffer
- > space available
- > MultiNet Server: FTP: I/O error %MULTINET-W-ENOBUFS, No buffer space available
- > -------------------------------------------------------------------------------
-
- > These problems were intermittent - they would appear then disappear. We
- > tried restarting the multinet server from the config/servers utility, and
- > killing the MULTINET_SERVER process and rerunning the START_MULTINET.COM
- > procedure, but after a while problems would return. It took a reboot to
- > clean things up, and we haven't seen the problem since.
-
- > Does anyone have any ideas what we need to tune to keep from doing a reboot
- > if this happens again?
-
- Next time it is happening open a support call with us at 408-427-4366
- or SERVICE@TGV.COM. Also, we recommend you upgrade to MultiNet V3.1;
- V3.0 is no longer supported.
-
- The first thing we'll need is to see the output from a MULTINET SHOW/BUF
- command to see if it has logged any "requests for memory denied".
-
- If it HAS, then the problem is either too low a value of SPTREQ or that
- you're machine is so busy that the default buffering is not enough (a very
- rare case, in which we'll give you a patch to raise it).
-
- If it HAS NOT, then the problem is most likely a wedged Ethernet
- controller which is refusing to transmit packets.
-
- In either case, we need to see the problem while it is happening.
-
- Ken
-
-