> Thanks for your reply. I have modified the multipath.conf file as you > suggested and restart multipathd. But still I am getting the same > result.
sorry, wasn't sure if that would fix it, but thought it was worth trying. Not sure what the next troubleshooting step would be. If powerpath is enabled on the same box is it preventing multipath from seeing it? > > Which id are you referring to? > > If its the pseudo name the mpathX is what you are looking > > for. The device > > itself would be like /dev/mapper/mpathX > > > No , i am not referring the pseudoname. I am referring the device id, which > is the identification of the lun. Without that id, when the nodes > are in cluster , how we will identify the shared luns? Okay... i think the equivalent with multipath is going to be you WWID. On the systems where we have multipath mass deployed we actually turned off friendly names and just rely strictly on the WWID for both identification and device name. This names the device to be unique, consistent and still exist on multiple systems. You can do this by commenting out the use_friendly_names line. We thought about trying to generate ids with udev rules (you might look into that), but realized it was to much work and required us to come up with something when the WWIDs worked just fine for us. > May be as I am habituated with PowerPath, so I am little bit > confused to understand the technique of the multipath. Its been a few years since I used multipath, and I didn't use it to the level you are so I doubt I'd be much more help understanding the differences. > If you give me some pointers , that will be really appreciated. > Please give me some idea , if we have two cluster nodes and we are > using some shared disks. In that case how we will identify/ > configure the shared luns in multipath? You might also go look up the "best practices" white paper from your vendor (either the san or the hba vendor). I know a few of them distribute recommended multipath.conf files. I would also verify that powerpath isn't interfering with your mapping. -greg _______________________________________________ rhelv5-list mailing list [email protected] https://www.redhat.com/mailman/listinfo/rhelv5-list
