Bug 91104 - autofs init script failing cause of daemon options
|
Bug#:
91104
|
Product: Gentoo Linux
|
Version: unspecified
|
Platform: All
|
|
OS/Version: Linux
|
Status: RESOLVED
|
Severity: major
|
Priority: P2
|
|
Resolution: FIXED
|
Assigned To: stefaan@gentoo.org
|
Reported By: wintered@gmail.com
|
|
Component: Applications
|
|
|
URL:
|
|
Summary: autofs init script failing cause of daemon options
|
|
Keywords:
|
|
Status Whiteboard:
|
|
Opened: 2005-05-01 20:38 0000
|
In version autofs-4.1.3-r4.ebuild, the init scripts were revamped, and now it
is unable to parse my daemon options.
The daemon options are: "-ro,soft,intr", the 't' in soft is intepreted as a
timeout flag with options ",intr".
Reproducible: Always
Steps to Reproduce:
1. Set daemon options to "-ro,soft,intr" for one of the mounts
2. Start the autofs init
Actual Results:
/home/usr/sbin/automount: option -t requires a numeric argument, got ,intr
* Failure on /home
It seems to come down to the command getopt not doing the expected
actually, --timeout 60 was added as a default setting in the /etc/conf.d/autofs
file. The new init script seems to not play well with options in the NIS map
or flat file if you start them with a '-' like normal. If you remove the dash,
it works fine. With the dash in place, it does weird things. Example
auto.master entry:
/apps/public auto.appspublic -rw,intr,nobrowse
Command that /etc/init.d/autofs tries to run:
/usr/sbin/automount --timeout r,nobrowse --pid-file
/var/run/autofs.apps.public.pid /apps/public yp auto.appspublic
-rw,intr,nobrowse
Yeah I noticed that removing the '-' makes it all work. I did a search and all
example include this flag though. Maybe something to do with the some of the
linux clients using autofs 3.
so, why we can't use ${options/-/} in getopt_input as workaround?
sounds good to me, who has access to make the change?
I'm working on a research group where we have a mixure of Debian, Ubuntu and
Gentoo machines (among other) and we are using YP + autofs.
With autofs-4.1.3-r4 I have the same problem you have described in the bug. I
have modified /etc/init.d/autofs taking into account the advice from Max
Loparyev (comment #3), and now it is working. The only modification was adding
options=${options/-/} to getmounts function, just inside the while loop.
I'm having the same problem after emerging uD world. Strange how an app with
this bug is put in portage as stable. I'm using NIS + NFS
Oh well,
I have:
# ypcat -k auto.master
/home auto.home -t 60 -rw,soft,intr,nosuid,rsize=8192,wsize=8192
/net auto.net -t 60 -rw,soft,intr,nosuid,rsize=8192,wsize=8192
# /etc/init.d/autofs stats
Configured Mount Points:
------------------------
/usr/sbin/automount --timeout ,intr,nosuid,rsize=8192,wsize=8192 --pid-file
/var/run/autofs.home.pid /home yp auto.home
-rw,soft,intr,nosuid,rsize=8192,wsize=8192
/usr/sbin/automount --timeout ,intr,nosuid,rsize=8192,wsize=8192 --pid-file
/var/run/autofs.net.pid /net yp auto.net -rw,soft,intr,nosuid,rsize=8192,wsize=8192
Active Mount Points:
--------------------
Which results in the numerical argument error to automount. The problem is I
cannot find a workaround to this. Someone said to "use ${options/-/} in
getopt_input" but I cannot find getopt_input. And I cannot remove '-' since I
don't know where is the config files. I tried to find everywhere in my maching
but couldn't find it. Occured to me it could be information provided by the
server in which case I cannot change. Any ideas?
*** Bug 145056 has been marked as a duplicate of this bug. ***
Someone please make a new revision and put it into portage asap! ${options/-/}
in getopt_input works well.
New maintainer, net-fs/autofs-5.0.3-r1 is now in the tree.
Is there still a demand for this? Or can we all just as well upgrade to
autofs-5 (which I guess doesn't have this problem, as the init script is, again
:), completely new)
It is a super simple fix, so I don't see why it can't be done. It is sad that
this languished for so long.
(In reply to comment #11)
> It is a super simple fix, so I don't see why it can't be done. It is sad that
> this languished for so long.
Sad is probably not the word I would use :)
(In reply to comment #12)
> Sad is probably not the word I would use :)
You are always welcome to spend some of your own free time at fixing bugs,
becoming a gentoo developer is an easy enough process...
Fixed in net-fs/autofs-4.1.3-r9