gomd-devel
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[gomd-devel] Re: <CW_TOOLS> new release (with oMFS checks)


From: Matthias Rechenburg
Subject: [gomd-devel] Re: <CW_TOOLS> new release (with oMFS checks)
Date: Sun, 24 Aug 2003 12:37:09 +0200
User-agent: KMail/1.4.3

Ciao JP,

On Sonntag 24 August 2003 11:03, Gian Paolo Ghilardi wrote:
> Hi Matt.
>
> "cw_tools" now have two new functions to check if oMFS is configured and
> started before doing any operation on files/directories.
>
> Checkout the new cw_tools release... :)
> I need comments/criticisms/hints...

ok, reviewed them. here are my comments ;)

in isOMFSConfigured i would prefer to grep in /proc/filesystems to
check if oMFS is configured/supported by the kernel instead of the fstab.
The user might have mounted it manually by some other script.
Also grep for mfs in df -h. Better as to grep for the mount-point 
from the fstab e.g. If "df -h | grep mfs" is successfull then you can awk 
the mount-point from it.

in isOMFSEnabled i would prefer to check if oMFS is mounted and
if not just create a temporary mount e.g. /tmp/mfs, mount oMFS on it,
install the gomd and then umount it again.
The user might not want to use oMFS but we can use it for the
cluster-wide installation.

... just my mind.

>
> Thanks.
>
> <rejected>

thank you ;))

have fun,

Matt

>
>
> ----- Original Message -----
> From: "Matthias Rechenburg" <address@hidden>
> To: "Gian Paolo Ghilardi" <address@hidden>; "gomd developers mailing list"
> <address@hidden>
> Cc: "Roel Baardman" <address@hidden>; "Mirko Caserta"
> <address@hidden>
> Sent: Sunday, August 24, 2003 9:42 AM
> Subject: Re: <CW_TOOLS> cluster-wide (un)installation for gomd/libgomd
>
> > Hey JP, Matt again,
> >
> > a small feature request for those cw_tools ;)
> >
> > Since all those scripts are based on oMFS there are basically
> > (about) 3, ..eh 4 situations :
> > - oMFS is enabled and mounted
> > - oMFS is enabled but not mounted
> > - oMFS is disabled
> > - running a non openMosix kernel -> no oMFS
> >
> > ....so to my mind those scripts should first check if it is an
> > openMosix kernel. Then grep in /proc/filesystems if oMFS
> > is supported by the kernel. After that they should "find" the
> > oMFS mount-point
> > (since it can be mounted everywhere, not only under /mfs)
> > e.g. by grepping the df output.
> > If oMFS is supported but not mounted the scripts also can create a
>
> temporary
>
> > mount-point, mount it, install gomd clusterwide and umount it again
> > (since the user decides if he/she wants to use oMFS or not)
> >
> > ....... just to make this nice feature more robust ;)))
> >
> > enjoy,
> >
> > Matt

-- 
E-mail  :  address@hidden
www     : http://www.openmosixview.com
an openMosix-cluster management GUI

for X in $(seq 1 50); do (cat /dev/zero |netcat localhost 9 &); done





reply via email to

[Prev in Thread] Current Thread [Next in Thread]