[gentoo-amd64] hald randomly unmounting partitions

[gentoo-amd64] hald randomly unmounting partitions

[Paul Colquhoun]

For a few months, I have been having problems where some partitions are 
being unmounted at odd intervals.

I wrote a script wrapper around the umount command, and have found that 
it is being done by a script that is part of the hald setup.

Has anybody else seen this?

The affected partitions are ones that usually don't have any open files, 
but only because the other umount's are failing, the script is running 
through *ALL* my partitions and trying to unmount them.

The script in question is:

/usr/share/hal/scripts/hal-system-storage-unmount


The umount wrapper is showing this activity:

Thu Jan 18 18:47:11 EST 2007 umount called with parameters = '/' : 
Parent PID = 20125
root     20125 20124  0 18:47 ?        
00:00:00 /bin/bash /usr/share/hal/scripts/hal-system-storage-unmount

Thu Jan 18 18:47:11 EST 2007 umount called with parameters = '/var' : 
Parent PID = 20134
root     20134 20133  0 18:47 ?        
00:00:00 /bin/bash /usr/share/hal/scripts/hal-system-storage-unmount

Thu Jan 18 18:47:11 EST 2007 umount called with parameters = '/usr' : 
Parent PID = 20143
root     20143 20142  0 18:47 ?        
00:00:00 /bin/bash /usr/share/hal/scripts/hal-system-storage-unmount

Thu Jan 18 18:47:12 EST 2007 umount called with parameters = '/data' : 
Parent PID = 20152
root     20152 20151  0 18:47 ?        
00:00:00 /bin/bash /usr/share/hal/scripts/hal-system-storage-unmount

Thu Jan 18 18:47:12 EST 2007 umount called with parameters 
= '/var/spool/news' : Parent PID = 20172
root     20172 20171  0 18:47 ?        
00:00:00 /bin/bash /usr/share/hal/scripts/hal-system-storage-unmount

Thu Jan 18 18:47:12 EST 2007 umount called with parameters 
= '/usr/portage' : Parent PID = 20186
root     20186 20185  0 18:47 ?        
00:00:00 /bin/bash /usr/share/hal/scripts/hal-system-storage-unmount

Thu Jan 18 18:47:12 EST 2007 umount called with parameters = '/home' : 
Parent PID = 20198
root     20198 20197  0 18:47 ?        
00:00:00 /bin/bash /usr/share/hal/scripts/hal-system-storage-unmount


/data  /var/spool/news  &  /usr/portage  are the partitions that are 
affected most.


Is there a way to mark these partitions so that hald will keep it's 
interfering mitts off them?


-- 
Reverend Paul Colquhoun, ULC.    http://andor.dropbear.id.au/~paulcol
     Asking for technical help in newsgroups?  Read this first:
        http://catb.org/~esr/faqs/smart-questions.html#intro
-- 
gentoo-amd64@gentoo.org mailing list

[gentoo-amd64] Re: hald randomly unmounting partitions

[Duncan]

Paul Colquhoun <paulcol@andor.dropbear.id.au> posted
200701182302.59291.paulcol@andor.dropbear.id.au, excerpted below, on  Thu,
18 Jan 2007 23:02:59 +1100:

> For a few months, I have been having problems where some partitions are
> being unmounted at odd intervals.

I waited to see if someone with more knowledge on the subject would
post a reply, but I don't see any, so hoping the following is helpful
despite my limited knowledge on the subject...

hal-0.5.7.1-r3 was released yesterday (the day you posted), with a
suspend/hibernate bugfix and a pci-ids bugfix, according to the changelog.
That doesn't seem to be directly your issue, but with a bit of luck...

Beyond that, just looking around hal doesn't seem to have a configuration
of its own altho there are some examples in /usr/share/doc/hal-<version
that may or may not be helpful.  They looked interesting, but I didn't
see any info on where they should be put to activate them if one did want
to run them.  I'm not sure that's the issue, however.

If it were me the next place I'd check would be my KDE or GNOME removable
drive configuration. If you thought you were setting it up to manage your
CD/DVD/UDB devices and it got misconfigured (either as merged or by you)
to treat /all/ mounts as removable...  Also, those examples in the
documentation mentioned above reminded me an awful lot of the udev
configuration stuff, so I'd take a peak at that and see if anything out of
place jumped out at me.

You already mentioned using a wrapper for troubleshooting, which would
imply you're decently experienced at it.  What about using that wrapper to
pause the thing, then doing a ps --forest (or use a GUI tool such as
ksysguard to get the process tree), in ordered to figure out what's
invoking it?  At least then you'd have a better idea what you might need
to reconfigure or disable or whatever, to get it to quit trying to unmount
your main partitions!

Hopefully that's of /some/ help, anyway.  As I said, I hoped someone with
a bit more hal knowledge would reply, but I don't see anything yet, and I
/did/ just see the hal update and remembered seeing your post, so decided
to mention it at least.

-- 
Duncan - List replies preferred.   No HTML msgs.
"Every nonfree program has a lord, a master --
and if you use the program, he is your master."  Richard Stallman

-- 
gentoo-amd64@gentoo.org mailing list