Eskimo North


          [Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

          Mail and other weird upgrade related problems...


          • To: outages-list@eskimo.com, eskimo-announce@eskimo.com, Jeff@WINCHELL.COM, ericj@eskimo.com, evol@eskimo.com, snocat@eskimo.com
          • Subject: Mail and other weird upgrade related problems...
          • From: Robert Dinse <nanook@eskimo.com>
          • Date: Sun, 1 Aug 1999 08:06:30 -0700 (PDT)
          • Resent-Date: Sun, 1 Aug 1999 08:06:37 -0700
          • Resent-From: eskimo-announce@eskimo.com
          • Resent-Message-ID: <"rQl2c1.0.h_7.y96ft"@mx1>
          • Resent-Sender: eskimo-announce-request@eskimo.com

          
               Had some strange problems after the upgrade.
          
               A number of strange problems occured, most related to slow connections to
          mail or connections timing out, but there were other spurious problems as well.
          Looking at logs, I saw random name server failures, but testing our name
          servers (from eskimo) didn't reveal any problems.  The mail machine itself
          showed no signs of ill-health. 
          
               Finally (about 7am after being up all night), I re-arranged the name
          servers that mail uses so that ns1.eskimo.com (which is a virtual interface on
          eskimo presently) wasn't first, and then I could connect to it right away. 
          
               I tried pinging ns1 from mail and couldn't.
          
               I did an arp -a and noticed the ethernet address was incorrect.
          
               Well, I've run into this problem before where SunOS wants to hang on to
          old arp entries after hardware is changed, and it's a real pain in a sizeable
          network because you can get it cleared from one machine but another machine
          with a bad entry will respond to an arp request with it's bad entry polluting
          the arp table of a machine you just clearned out. 
          
               No matter what I did I couldn't get rid of this one bad entry, was driving
          me crazy, even rebooted the switching hub and router thinking maybe they were
          hanging onto it. 
          
               Then I found it was only hosed for the virtual interface, but not the main
          address.  I looked at the initialization file for virtual interfaces and it was
          hard-coded into the file.  Fixed that, rebooted, cleared all the bad entires,
          and now everything appears to be back to normal. 
          
          
          

          • Prev by Date: Eskimo Server Upgrade - Outage July 31, 1999 1am-3am
          • Next by Date: Re: User Meetings
          • Prev by thread: Re: User Meetings
          • Next by thread: Eskimo Server Upgrade - Outage July 31, 1999 1am-3am
          • Index(es):
            • Date
            • Thread