CX300 disk allocation question

[nyck]

I have got a V890 running solaris9 which is attached to a Clariion CX300 and I have just allocated some LUNS from the CX300 to this V890. After the boot -r I have run format and I get the following output:-

20. emcpower0g <DGC-RAID5-0219 cyl 40958 alt 2 hd 256 sec 10>
/pseudo/emcp@0
21. emcpower1g <DGC-RAID5-0219 cyl 40958 alt 2 hd 256 sec 10>
/pseudo/emcp@1
22. emcpower2a <DGC-RAID5-0219 cyl 32766 alt 2 hd 32 sec 10>
/pseudo/emcp@2
23. emcpower3a <DGC-RAID5-0219 cyl 49150 alt 2 hd 64 sec 10>
/pseudo/emcp@3
24. emcpower4c <DGC-RAID5-0219 cyl 49150 alt 2 hd 64 sec 10>
/pseudo/emcp@4
25. emcpower5e <DGC-RAID5-0219 cyl 49150 alt 2 hd 64 sec 10>
/pseudo/emcp@5
26. emcpower6h <DGC-RAID5-0219 cyl 49150 alt 2 hd 64 sec 10>

I would of thought that these powerpath devices would of been emcpower0a,emcpower1a,emcpower2a etc but as you can see from the above output this is not the case. Can someone tell the reason for this?

All the mount points I have running from these devices work fine!

 

[Annihilannic]

Unfortunately the device enumeration order used by EMC PowerPath is not necessarily the same as that used by the OS, so you can end up with situations like that. As you say, it does no harm, but it sure looks messy.

It can be further complicated if you are adding new storage to a system that already has some allocated PowerPath devices; if you really like to be neat you can remove or rename the PowerPath device database (location depends on your version of PowerPath, but I think it's something like /etc/emcp_devicesdb.dat) and do another reconfigure reboot to force the devices to be enumerated again, but remember it may have implications if you refer to these devices directly anywhere else, e.g. /etc/vfstab as the devices will probably have new names after the reboot.

Annihilannic.