|Summary:||wrong chkconfig stuff|
|Product:||[Retired] Red Hat Raw Hide||Reporter:||dharris|
|Component:||bind||Assignee:||Cristian Gafton <gafton>|
|Status:||CLOSED NEXTRELEASE||QA Contact:|
|Fixed In Version:||Doc Type:||Bug Fix|
|Doc Text:||Story Points:||---|
|Last Closed:||1999-07-26 19:01:19 UTC||Type:||---|
|oVirt Team:||---||RHEL 7.3 requirements from Atomic Host:|
|Cloudforms Team:||---||Target Upstream Version:|
Description dharris 1999-07-24 01:35:24 UTC
Comment 1 Bill Nottingham 1999-07-26 17:04:59 UTC
Nope. It's a change in policy; we don't want to start named automatically if installed anymore. chkconfig --add shouldn't change existing links, though (if it does, that is a bug). Of course, now I look at the package and see that the scripts that it includes don't mesh with this. Will fix.
Comment 2 dharris 1999-07-26 18:41:59 UTC
The "chkconfig: - 55 45" line works to make named install by default as non-running, but the problem is that I can't enable the service using chkconfig... that line basically "short-circuits" chkconfig's definition of the service running so that "chkconfig --add named" and "chkconfig --del named" have the same effect. To enable the service I have to manually create the symlinks. The whole purpose of chkconfig was to keep me from having to manually create these links. Wouldn't it just be better to leave the "chkconfig: 345 55 45" line alone and change the "chkconfig --add" line in the postinstall script to "chkconfig --del" to implement this policy change?
Comment 3 Bill Nottingham 1999-07-26 18:47:59 UTC
Actually, the chkconfig behavior works like this: chkconfig --add adds the default links specified in the init file. chkconfig --del removes *all* links. To enable a service using chkconfig, regardless of the initfile defaults, do something like: chkconfig --level 345 named on (The --level argument is a list of runlevels. --level <foo> is optional.)
Comment 4 dharris 1999-07-26 18:57:59 UTC
Oh, that's funny. I've been using chkconfig wrong all this time and I've not gotten bitten yet. :-) Sorry about the erroneous bug report.
Comment 5 Bill Nottingham 1999-07-26 19:01:59 UTC
That's OK. At least it caused me to notice the bind chkconfig & links disagreeing. :)