On Sun, Sep 25, 2005 at 04:15:46AM +0200, Carsten Hey wrote: > But i just remembered one thing that might be important. I added the option > -d 10 to evms_activate in the init script and I was able to mount something. > After removing -d 10 again I had the old behavior.
-d 10 is an invalid switch (-d goes from 0 to 9), so I'm not quite sure what the effects of this should be. :-) Anyhow, I think what you're seeing is the intended behaviour, although you might be misinterpreting it. By default, EVMS tries to discover all partitions (both EVMS and non-EVMS; ie. so-called “compatibility volumes”) and add volumes in /dev/evms for them at startup. Due to bd_claim (see the link I pasted) it can't do this for /dev/hda*, since the kernel has already locked /dev/hda before evms_activate is run. However, it can do it for /dev/hdc, since you don't have anything mounted there, so you get /dev/evms/hdc1 etc. as compatibility volumes. But at this point, the kernel locks hdc (since EVMS is using it), and you can't mount /dev/hdc1. There are three solutions to this: - Exclude /dev/hdc in your EVMS configuration. - Apply the bd_claim patch. - Mount /dev/evms/hdc1 instead of /dev/hdc1. In any case, this is AFAICS not a bug. /* Steinar */ -- Homepage: http://www.sesse.net/ -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
