On Fri, Feb 15, 2008 at 04:22:07PM -0000, Rogan, Julian wrote: > Hi, > I am using multipath for the first time with little joy. <snip> > based on some other documents I made the following changes to > /etc/multipath.conf You can probably leave the entire /etc/multipath.conf commented out. Everything should just autodetect correctly. > I ran > multipath -v3 to recreate I get the following: <snip> > device-mapper ioctl cmd 14 failed: Invalid argument > message 3600508b300930230a7495ae1b9500074 0 switch_group 1 > failed > > Note sure if "path sdX not found in pathvec" relevant No, that's supposed to be there. > anyway no mpaths have been created. > > I hope there is someone out there who can give me some help with this. No chance you can update to a more recent version of RHEL4? The device-mapper message means the command you send to device-mapper didn't work. However, it looks completely sensible. What version of device-mapper are you using? # rpm -q device-mapper I'm guessing that you will see something like device-mapper-1.01.04-1.0.RHEL4 device-mapper-multipath-0.4.5-6.0.RHEL4 needs device-mapper-1.02.* Look at things from my side, it definitely seems like RHEL 4.2 says it uses device-mapper-1.01, which doesn't make any sense, since device-mapper-multipath-0.4.5-6.0.RHEL4 was definitely compiled against device-mapper-1.02, and I don't believe that the can both 1.01 and 1.02 can coexist on a system. The easy solution is to upgrade to at least RHEL 4.3 If that's impossible, you can try to upgrade device-mapper to device-mapper-1.02.02-1.0.RHEL4 or later. This might cause all sorts of dependancy issues. It may be easier to downgrade device-mapper-multipath to device-mapper-multipath-0.4.5-5.2.RHEL4 I'm not sure why this incompatibility exists now, since it obviously didn't when RHEL 4.2 was originally released. My best guess is that device-mapper-multipath-0.4.5-6.0.RHEL4 didn't really get released until RHEL 4.3, but that's just a guess. Of course, if you already have device-mapper-1.02.* on your system, and are still seeing these issues, then something else is going on. -Ben > with regards, > Julian Rogan > -- > dm-devel mailing list > dm-devel@xxxxxxxxxx > https://www.redhat.com/mailman/listinfo/dm-devel -- dm-devel mailing list dm-devel@xxxxxxxxxx https://www.redhat.com/mailman/listinfo/dm-devel
