Bug 507763 - ypbind shouldn't try to start until network resources are ready
Summary: ypbind shouldn't try to start until network resources are ready
Keywords:
Status: CLOSED DUPLICATE of bug 480096
Alias: None
Product: Fedora
Classification: Fedora
Component: yp-tools
Version: 11
Hardware: All
OS: Linux
low
medium
Target Milestone: ---
Assignee: Karel Klíč
QA Contact: Fedora Extras Quality Assurance
URL:
Whiteboard:
Depends On:
Blocks:
TreeView+ depends on / blocked
 
Reported: 2009-06-24 05:59 UTC by Braden McDaniel
Modified: 2013-03-03 22:59 UTC (History)
6 users (show)

Fixed In Version:
Doc Type: Bug Fix
Doc Text:
Clone Of:
Environment:
Last Closed: 2010-04-15 15:47:15 UTC
Type: ---
Embargoed:


Attachments (Terms of Use)
/var/log/messages with NetworkManager enabled (293.79 KB, text/plain)
2009-09-17 07:13 UTC, Braden McDaniel
no flags Details

Description Braden McDaniel 2009-06-24 05:59:51 UTC
Description of problem:
ypbind fails at boot because it cannot find the NIS server when using NetworkManager.  After booting, starting ypbind works fine.  If NetworkManager is disabled and network is enabled, ypbind will successfully start at boot.

Version-Release number of selected component (if applicable):
0.7.1-4.git20090414.fc11

How reproducible:
Consistently (on multiple machines).

Comment 1 Niels Haase 2009-06-24 08:55:25 UTC
Thanks for filling this bug. To get closer to your problem, can you pleas attach your /var/log/messages after boot with NM and also with s-c-n enabled? Thank you.

-- 
Fedora Bugzappers volunteer triage team
https://fedoraproject.org/wiki/BugZappers

Comment 2 Braden McDaniel 2009-09-17 07:13:25 UTC
Created attachment 361435 [details]
/var/log/messages with NetworkManager enabled

Comment 3 Braden McDaniel 2009-09-20 08:54:35 UTC
When I restart NetworkManager, it looks like this:

[root@bolt ~]# service NetworkManager restart
Stopping NetworkManager daemon:                            [  OK  ]
Setting network parameters...                              [  OK  ]
Starting NetworkManager daemon:                            [  OK  ]
[root@bolt ~]# r8169: eth0: link up

Note the last message.  This occurs a second or two after service has returned.  It looks like NetworkManager is bringing up eth0 asynchronously--returning with "OK" before the network is necessarily available.  And I think that if it's *not* available when ypbind starts (which it does immediately afterward during startup), ypbind fails.

It certainly seems like NetworkManager shouldn't be returning with "OK" until the network is available for use.

Comment 4 Dan Williams 2009-09-21 04:16:06 UTC
NetworkManager is designed to bring interfaces up and down asynchronously; we need to penalize (or fix) the services that can't handle that instead of penalizing all services.

There are two options:

1) Fix the ypbind initscript to block until the network is actually available

2) set NETWORKWAIT=yes in your /etc/sysconfig/network file

Does #2 work for you?

Comment 5 Braden McDaniel 2009-09-21 06:27:50 UTC
Yes; adding NETWORKWAIT=yes works around the problem.

Comment 6 Dan Williams 2009-11-12 01:27:39 UTC
To fix this in yp-tools, maybe a dispatcher script would work to restart ypbind when the network situation changes.  See initscripts (which drops /etc/NetworkManager/dispatcher.d/05-netfs onto the system) and 'man NetworkManager' for more information on dispatcher scripts.

Comment 7 Dan Williams 2009-11-12 01:28:15 UTC
(last comment more for yp-tools maintainer than for Braden)

Comment 8 Karel Klíč 2010-04-15 15:47:15 UTC
Hi, ypbind should be able to handle the NetworkManager async startup now. 

When started, ypbind handles unavailable network just fine AFAIK.

*** This bug has been marked as a duplicate of bug 480096 ***


Note You need to log in before you can comment on or make changes to this bug.