Premature EOF - Really A Memory Problem?

[Danny Mayer]

At 04:33 PM 10/2/01, Kevin Vaughn wrote:
>I'm getting the following messages in my named.run log:
>
>*****
>02-Oct-2001 14:39:58.000 security: info: approved AXFR from [10.5.90.4].3808
>for "pccatest.com"
>02-Oct-2001 14:39:58.000 xfer-out: info: zone transfer (AXFR) of
>"pccatest.com" (IN) to [10.5.90.4].3808
>*****
>
>I also get the following in the event log:
>
>*****
>premature EOF, fetching "pccatest.com"
>*****
>
>All 120 of my reverse lookup files are transferring.  My zone file which
>contains all of the A records is not transferring.  There are around 1400
>hosts in that file.
>
>I have read all of the archives from this list and everywhere else I can
>find.  Danny Mayer has said time and time again that this is a memory
>problem.  My pagefile size is set to 1024MB, I have 64MB of physical memory.
>I have no errors in the event log for the primary server.  The only errors I
>have on the slave are the ones listing the premature EOF error.  Is this
>problem really a memory problem or could it be something else?  It just
>doesn't make sense to me that I would be running out of memory.  I don't see
>any spikes in memory during replication.  I can watch the resource monitor
>and come to that conclusion. 

Which machine are you monitoring?  It's the master, not the slave that causing
the problem.

>  If I remove half of the records from the zone
>file then the file replicates with no errors.  If I put the records back in
>it replication fails once again.  I am running BIND 8.2.5-REL.

         Danny