My company is setting up an RHEL server to test the performance of Oracle on Linux. This is our first experience with Linux in the Enterprise and we need a little help.

We installe RHEL on a PowerEdge server with a QLogic fiber card connected to our EMC storage device and carved out a 100 GB partition on the SAN. Now we need to mount that partition somewhere, but we are stuck.

Following a reboot we have a new device /dev/sdb that we assume is the new 'disk' but when we try to mount it we get an error 'can't read superblock'.

Could someone please point me to some step by step instructions for configuring and mounting the new partition.

I have been using Linux at home for years and run a RH web/data server, but this is our/my first experience trying to connect a Linux server to the SAN. Any help would be greatly appreciated. I am committed to making this project a success in the hope that my company embraces Linux.

this *is* only a trial right??

the device you'll be presenting is seen as a block device, e.g. a hard drive. you would need to partition it as you speculate. use fdisk to partition it and then format the relevant partitions, e.g. "mke2fs -j /dev/sda1"

Yes, this is only a trial, or more precisely a proof of concept.

We tried to partition it using fdisk, but got an error, I can't recall the error off the top of my head, I will try again at work on Monday and get the error message.

I ran the following command:
fdisk /dev/sbd
and got:
Unable to read /dev/sdb

does /dev/sdb actually exist? ´ls -l /dev/sdb´

I use the gui system settings --> Logical Volume manager. You'll need to pvcreate the partition.

Forgot to ask are you using power path? If so, you will need to intialize the /dev/emcpowera. Here are my cryptic notes. It failed for me but it created a device that you'll need to pvcreate.

pvcreate command failed. Command attempted: "/usr/sbin/pvcreate -M 2 /dev/emcpowera1" - System Error Message: Device /dev/emcpowera1 not found (or ignored by filtering).


The Fix: Go to the command line prompt

1. Cd /dev
2. ll emcpower(letter/number) example emcpowera1
3. pvcreate –M2 /dev/emcpower(letter/number)

if you do have powerpath, you'll need to install EMCpower.LINUX - the rpm can be bound in powerlink.emc.com
if you have it installed, run:

powermt display dev=all

to list the LUNS in your system. if you see the emcpower devices listed and the corresponding LUN numbers, you need to rescan your fc_hosts - if you have restarted your machine, this should be done at startup.

to partition the disk, locate the emcpower devices (ie /dev/emcpowerX). and run fdisk /dev/emcpowerX.

for devices like this, consider using LVM2.

Firstly, I would recommend you to use emcgrab (Powermt RPM as pointed out by one of the senior member) software since RHEL 2.1 is having some issues with pseudo device bindings.

cat /proc/scsi/qlaXXXX/0

this will give you the idea whether the LUN has been properly presented to your server or not.

/dev/emcpowera should be the first pseudo device which binds itself to /dev/sda or /dev/sdb depending upon emcpower.conf file.

You can fdisk the same without any issues.

powermt display dev=all| grep emcpower will display all the emc pseudo devices which have been automatically attached with /dev/sd* device.

Remember /dev/emcpower* are indeed the block devices but they are actually attached to /dev/sd* device.

Hope this helps!!
Rahul Khare

Thanks for all the input, I will let you know how things go.

To answer Chris' question, yes /dev/sbd does exist.

# pvcreate -ff /dev/sdb /dev/sdc
Device /dev/sdb not found (or ignored by filtering).
Device /dev/sdc not found (or ignored by filtering).


Solution :

Vi the /etc/lvm/lvm.conf and search for name line "filter" . Got to end of the line and remove the string "r/.*/" , and save it .


Before editing:

filter = [ "r|^/dev/mapper/.*-lv.*|", "a|^/dev/mapper/.*|", "r/.*/" ]



Should look like

filter = [ "r|^/dev/mapper/.*-lv.*|", "a|^/dev/mapper/.*|"]


Then run the pvcreate on the disks ;

Just go to http://www.doxer.org/learn-linux/emcgrab-download-url/ to have emcgrab download.

R u sure ur backend has SATA disk? Which array u r using? Confirm first with naviseccli. It seems that u r not using multipath software. Make sure first that Zoning is proper. LUN masking is working properly. If everything goes well, u need to scan the devices from ur RHEL host. I can't remember the command. Search in EMC site.