Re: Risks of Frequent yppush in distributed environment?

From: Alek O. Komarnitsky <alek_at_mail1.ast.lmco.com>
Date: 1997/07/29
Message-ID: <5rl6kl$922$1_at_mail1.ast.lmco.com>#1/1


In article <5rl0mv$a9t_at_dfw-ixnews9.ix.netcom.com>, Errol Back-Cunningham <ebc_at_ix.netcom.com> wrote:
>
> Came across a site that uses HP-UX and has a distributed database
> environment - uses Sybase, Oracle and Informix, thousands of users,
> numerous platforms and does a yppush every five minutes.
>
> Does anyone have any idea what (if any) risks are associated with
> this? I'm of the opinion that any changes to a live production
> environment are inherently risky, and doing a yppush of large
> files every 5 minutes seems to me to be excessive.
>
> Any idea what effects we might see when we encounter network
> spikes or problems in mid-push?
>
> Errol

FWIW: Our site has ~1,000 workstations (mostly Sun & HP), and we currently have 4,103 password entries, 18,591 host entries, and 44,103 Email aliases (and 132 YP-Slaves ... probably a few too many there ... but we have a lot of subnets).

We don't push every 5 minutes ... but you can image that stuff get pushed all the time. We get the occasional "corrupt" map; where it just seems screwed up ... but this happens rarely and we just push back out and things are fine.

I would be curious to hear comments from other folks ...

alek Received on Tue Jul 29 1997 - 00:00:00 CEST

Original text of this message