[THIN] Re: fault tolerance print servers

  • From: "Berny Stapleton" <berny@xxxxxxxxxxxxxxxxx>
  • To: thin@xxxxxxxxxxxxx
  • Date: Mon, 9 Jun 2008 14:06:01 +0100

We are migrating our print servers into VMware at the moment, so if I can
move it onto a cluster I will. I need a clustered web server anyway, so two
birds one stone as far as I can see.....

2008/6/9 Berny Stapleton <berny@xxxxxxxxxxxxxxxxx>:

> Sorry to reply to my own post, but seeing Chad's post, that's what it was
> using, NLB under NT4. Point and print didn't work with that, it always
> prompted for drivers, same as Novell back in those days.
> 2008/6/9 Berny Stapleton <berny@xxxxxxxxxxxxxxxxx>:
> I know this is an old question, but I haven't looked at it in that long.
>> Did they get point and print working? It used to be (Under NT4) that you
>> could create a highly available print cluster but point and print didn't
>> work. I simply just haven't clustered a print server for that long and I
>> don't have a test cluster to play with at the second.
>> Berny
>> 2008/6/9 Adam Thompson <adwulf@xxxxxxxxx>:
>> 2008/6/9 Greg Reese <gareese@xxxxxxxxx>:
>>> > you could build two servers and build the same queues on each one.
>>>  Then use
>>> > an alias in DNS to point the users to one or the other.  Should the
>>> main one
>>> > fail, change the alias in DNS to the other one and clients should keep
>>> > working.  it's not automatic unless you do some clever monitoring and
>>> > scripting.  And you might lose a couple print jobs in the tranisition
>>> during
>>> > a failure.
>>> >
>>> > there's a few holes in this but it is basic and easy.
>>> >
>>> You'll need to disable strictnamechecking to do this.
>>> Then assign a CNAME which can point to either server (eg PRINTSERV),
>>> point it at your preferred server initially (eg SERV1)
>>> Now, you need to have something which detects when the node has died,
>>> and runs something like dnscmd.exe /RecordDelete and dnscmd.exe
>>> /RecordAdd to modify the CNAME.
>>> It won't be instantaneous, as you'll have to wait for DNS to replicate
>>> around.  Also, all the jobs in the queue will appear to be lost, and
>>> you still have the problems of client DNS cache and what happens when
>>> the nodes can talk to the dns server but not each other (possibly
>>> worked around with a heartbeat NIC).
>>> Much simpler, if you have W2k or later to take a look at:
>>> http://download.microsoft.com/download/5/b/a/5bae5aa4-674f-42fb-b8cd-49e26bd29549/clusterfp.doc
>>> Creating Highly Available File and Printer Shares with Windows 2000
>>> Cluster Service
>>> or
>>> http://www.microsoft.com/windowsserver2003/techinfo/overview/availableprinter.mspx
>>> Creating and Configuring a Highly Available Print Server Under
>>> Microsoft Windows Server 2003 Using a Server Cluster
>>> --
>>> AdamT
>>> "At times one remains faithful to a cause only because its opponents
>>> do not cease to be insipid." - Nietzsche
>>> ************************************************
>>> For Archives, RSS, to Unsubscribe, Subscribe or
>>> set Digest or Vacation mode use the below link:
>>> http://www.freelists.org/list/thin
>>> You can follow the Thin List on Twitter:
>>> http://twitter.com/thinlist
>>> Thin List disucssion is now available for reading
>>> in blog format at:
>>> http://thinmaillist.blogspot.com
>>> ************************************************

Other related posts: