Nsdc notify should process in parallel

NSD
1

Hi,

If you have many zones, and you have a few zones with a broken/bad nameserver,
then the notify process stalls using exponential backof, delaying all other
zones sequentially as well. Example:

[root@vault nsd]# nsdc notify
Sending notify to slave servers...
[1269815681] nsd-notify[22093]: warning: timeout (1 s) expired, retry notify to 193.110.157.23.
[1269815683] nsd-notify[22093]: warning: timeout (2 s) expired, retry notify to 193.110.157.23.
[1269815687] nsd-notify[22093]: warning: timeout (4 s) expired, retry notify to 193.110.157.23.
[1269815695] nsd-notify[22093]: warning: timeout (8 s) expired, retry notify to 193.110.157.23.
[1269815711] nsd-notify[22093]: warning: timeout (16 s) expired, retry notify to 193.110.157.23.
[1269815743] nsd-notify[22093]: warning: error: failed to send notify to 193.110.157.23.
[1269815744] nsd-notify[22116]: warning: timeout (1 s) expired, retry notify to 193.110.157.23.
[1269815746] nsd-notify[22116]: warning: timeout (2 s) expired, retry notify to 193.110.157.23.
[1269815750] nsd-notify[22116]: warning: timeout (4 s) expired, retry notify to 193.110.157.23.
[1269815758] nsd-notify[22116]: warning: timeout (8 s) expired, retry notify to 193.110.157.23.
[1269815774] nsd-notify[22116]: warning: timeout (16 s) expired, retry notify to 193.110.157.23.
[1269815806] nsd-notify[22116]: warning: error: failed to send notify to 193.110.157.23.

[ many more times repeated ]

The real fix I guess is to process zones parallel instead of serial.

Paul

2

Hi Paul,

You are right that nsd-notify handles notifies sequentially and that
this delay can occur in specific situation.

However, note that NSD 3.x.x sends notifies automatically, implemented
non-sequentially. Also note that nsd-notify is deprecated since 3.x.x
and it is only still there because it is a useful aiding tool.

Best regards,

Matthijs

Paul Wouters wrote: