Summary: | Network services not starting: "... is scheduled to start when ..." | ||
---|---|---|---|
Product: | Gentoo Linux | Reporter: | Markus Peloquin <markus> |
Component: | [OLD] baselayout | Assignee: | Gentoo's Team for Core System packages <base-system> |
Status: | RESOLVED INVALID | ||
Severity: | normal | ||
Priority: | High | ||
Version: | 2006.1 | ||
Hardware: | x86 | ||
OS: | Linux | ||
Whiteboard: | |||
Package list: | Runtime testing required: | --- |
Description
Markus Peloquin
2006-11-10 11:46:08 UTC
One more thing is that this bug is similar to http://bugs.gentoo.org/show_bug.cgi?id=118801 . That bug does not really have a solution, though. Sounds like you need to set RC_STRICT_NET_CHECKING="lo" in /etc/conf.d/rc Or if you're feeling brave, try baselayout-1.13.0_alpha I tried the newest version of baselayout. It didn't fix the problem, and so far it only made things difficult (it is masked afterall). Also, and I don't want to spend too much time on this because I'm awfully busy, I don't think setting the services to start after net.lo will help. I would assume (and I'm not certain) that some of them need to have a real IP address when they start. For example, apache2 when starting without a real network cannot find the hostname so just uses 192.168.0.101. But then again, I suppose it's less of a hack than editing the init scripts. (In reply to comment #3) > For example, apache2 when starting without a real network cannot find the > hostname so just uses 192.168.0.101. But then again, I suppose it's less of a > hack than editing the init scripts. OK, so you really are using stuff that binds to IP's defined by other scripts. What you're seeing isn't a bug with baselayout. You've emerged netplugd/iplugd/wpa_supplicant/pppd to control the interface link. These services configure in the background which is why you're seeing the "scheduled to start" message. Well, when the link does come up then those services will get started. This is correct behaviour. You previously said > I remedied the situation by commenting out the "need net" line from the > initscripts that refuse to start, but I know this is not the correct solution. Which is the same thing as setting RC_STRICT_NET_CHECKING="lo" So what is the real error here? Is it that netplug/ifplugd/etc don't actually bring the link up? If so, what network card is it? Just noticed something. This is probably why I have to keep typing 'dhcpcd eth1' when I use wireless. $ ps -ef | grep ifplugd root 8159 1 0 15:53 ? 00:00:00 /usr/sbin/ifplugd --iface=eth0 So, it appears that ifplugd isn't starting on each interface. I tried adding 'ifplugd' to modules_eth1 in /etc/conf.d/net to no avail. No matter which network interface is loaded, ifplugd starts with net.eth0 but it never starts with net.eth1. I'm not sure, but this might be the issue. eth0 is like that one friend. You know, the one you hate. You have one piece of your pizza left, and he comes in and say, "oh thanks!" A-ha! I had in September switched from wireless-tools to wpa_supplicant because my school switched to WPA this year. My computer worked fine for a while. I recently found out that wpa_cli wrote to log, so I checked out syslog and found out that /etc/init.d/net.eth1 was failing to start because of an illegal module name 'mode'. There was an offending line in /etc/conf.d/wireless: config_EECS_WPA=( "mode managed" ) # 'EECS_WPA' is the ESSID at my school Okay, so nothing's wrong with baselayout. My network services start and everything now. Sorry if I caused any grief. |