High CPU Utilization on BIND 9.3 and dns_journal_compact failed: out of range error

[Ron Dutton]

Bind Users,

I was directed by Paul Vixie to this email address. Please assist here. Does a corrupted journal or dns_journal_compact ( ) spike CPU utilization for the named process ?


1631.        [bug]                         dns_journal_compact() could sometimes corrupt the journal [RT #11124]

Likewise, please answer my questions below.


Mark Andrews,

Thanks for the prompt reply.
It seems as thought the "dns_journal_compact failed: out of range" error only occurs on slave name servers which are
configured to receive notify messages from master name servers right before ixfr zone transfers occur.

My questions are :

1.    On the secondary itself, isn't the journal file the dynamic update log of new dynamic DNS updates received from the master?

2.    If we delete the .jnl file on the secondary won't new updates could be lost on this particular slave name server?

3.     What is an efficient way to purge .jnl files since we are using ixfr? Said another way, since the xfer of data is incremental
      changes are being pulled all the time. Thus, that could mean we would have to do multiple purges of the .jnl files on the slave.
      Please elaborate on this one.

4.    Do you suggest implementing use-ixfr no; in the named.conf and disable notify?


Please advise.


Regards,
Ron Dutton
Deutsche Bank
Network Engineering
IP Services


--

This e-mail may contain confidential and/or privileged information. If you are not the intended recipient (or have received this e-mail in error) please notify the sender immediately and destroy this e-mail. Any unauthorized copying, disclosure or distribution of the material in this e-mail is strictly forbidden.