Note: "permalinks" may not be as permanent as we would like,
direct links of old sources may well be a few messages off.
Steven wrote: Also, I believe you shoud add this in the script dmsetup mknodes ##maybe not this line vgscan --ignorelockingfailure vgchange -ay --ignorelockingfailure To make sure the volumes are scanned and activated. I am testing now to see if this will work when the drdb device is created over a linux partition (over hardware raid10) with no file system, I see no reasons why but since I could not find any recommendations about the right stacking of devices one has to experiment. Regards, > Leroy van Logchem wrote: >> Steven wrote: >>> Hi, >>> >>> I have LVM2 on top of DRBD and it seem to work fine(with the proper >>> LVM device filters: ), this is a file server. >>> I use it with with Xen, so I do NOT have the haresource configure to >>> mount any LV as these are brought up by Xen. >>> >>> The problem is that when a host is brought down the hot standby is >>> brought on-line BUT the vgchange command is NOT issued at the right >>> time or not at all: >>> *Secondary node* >>> [root at deimos lvm]# lvscan >>> inactive '/dev/DATA/VSI' [1.00 GB] inherit >>> inactive '/dev/DATA/IMG' [10.00 GB] inherit >>> >> Try adding this just after the drbddisk:: in haresources and before >> the mounting of filesystems. >> >> #!/bin/sh >> # >> # Startup script for Logical Volume Management >> # Arthur van Rooij @ WL|Delft Hydraulics july 2007 >> # >> ## >> >> start() >> { >> echo "Setting up Logical Volume Management" >> /sbin/vgscan >> /sbin/vgchange -a y >> sleep 5 >> } >> >> stop() >> { >> echo "Stopping Logical Volume Management" >> /sbin/vgchange -a n >> } >> >> # See how we were called. >> >> >> case "$1" in >> start) >> start >> ;; >> >> stop) >> stop >> ;; >> >> restart) >> stop >> sleep 2 >> start >> ;; >> *) >> echo "usage: $0 {start|stop|restart}" >> exit 1 >> ;; >> esac >> >> > Thanks a lot, that solved my issue perfectly the first shot. > > Regards, > > > -- Steven Dugway