ypservers file vs. ypservers map? 1

[z28sarefun]

OK, I know they aren't the same thing. So far I have:

If you don't have a ypservers file you must start up ypbind with a -broadcast option and have a NIS master or slave on the same subnet. Makes sense.

Now, if I have a ypservers file, are those the only servers that the machine will bind to? What is the role of the ypservers NIS map at that point? If I only have one machine in my ypservers file and it fails, will a machine fail over to a slave listed in the ypservers map?

Just about to replace our NIS master and I want to be sure how the map/file fit into the grand scheme.

Thanks,
Brian

 

[coffeysm]

In my experience if you specify a server in the ypservers file, the client will only try to bind to whats listed. So if your primary goes down without the secondardy listed it will hang. The ypservers map I believe is for the NIS server itself to know who to propagate the maps to.

 

[z28sarefun]

So I know that if I don't have a ypservers file the machine broadcasts for a NIS server/slave on the subnet. If there is more than one slave on a given subnet, will the machine broadcast again in case of a failure of the server it is bound to?

I'm trying to achieve failover without having to maintain a bunch of files.

 

[coffeysm]

Yes, it should resend out another broadcast if your primary server goes down.