Take the 2-minute tour ×
Server Fault is a question and answer site for system and network administrators. It's 100% free, no registration required.

I have problem with discovering my FC-connected CX-300 storage. Frankly speaking I'm complete novice in FibreChannel, so step by step explanation would be appreciated. My configuration consist of two IBM HS20 blades with RHEL5.4 on board and 2x Qlogic ISP2422-based 4Gb Fibre Channel HBAs on each blade. As a FC switch there are two Brocades built in BladeCenter Chassis, and finally there is EMC Clariion CX-300.

CX300, and Brocade switches should be configured properly, because they were working fine with previous configuration, which main defference was RHEL3 instead RHEL5.4

Below there is my output from several usefull commands:

#lspci | grep Fibre
06:01.0 FibreChannle: Qlogic Corp. ISP2422-based 4Gb Fibre Channel to PCI-X HBA (rev 02)
06:01.1 FibreChannle: Qlogic Corp. ISP2422-based 4Gb Fibre Channel to PCI-X HBA (rev 02)


#lsmod | grep qla
qla2xxx             1084741 0
scsi_transport_fc     37577 1 qla2xxx
scsi_mod             141717 10 scsi_dh,qla2xxx,sg,scsi_transport_fc,usb_storage,libata,mptspi,mptscsih,scsi_transport_spi,sd_mod

#cat /proc/scsi/scsi
Attached Devices:
Host: scsi0 Channel: 00 Id: 00 Lun: 00
  Vendor: LSILOGIC Model: 1030 IM    IM Rev: 1000
  Type:   Direct-Access                 ANSI SCSI revision: 02
Host: scsi0 Channel: 01 Id: 00 Lun: 00
  Vendor: IBM-ESXS Model: ST936701LC FN Rev: B418
  Type:   Direct-Access                 ANSI SCSI revision: 04
Host: scsi0 Channel: 01 Id: 00 Lun: 00
  Vendor: IBM-ESXS Model: ST936701LC FN Rev: B418
  Type:   Direct-Access                 ANSI SCSI revision: 04

I'd followed instructions from this site (editing /etc/multipath.conf), but i failed after multipath -ll - the output was empty.

Do you have any suggestions about discovering FC Connected LUNs in such configuration?

share|improve this question

2 Answers 2

This configuration is not specific to EMC.

  1. Get your storage properly presented by your storage admin. You can assist the storage admin by providing the WWPN (like a fiber channel mac address) of each of your fiber channel NICs to him. You can find them by running the following command:

    for i in $(ls -1 /sys/class/scsi_host/); do cat /sys/class/scsi_host/$i/lpfc_symbolic_name; done;

    or

    for i in $(ls -1 /sys/class/scsi_host/); do cat /sys/class/scsi_host/$i/device/fc_host\:$i/port_name; done;

  2. Rescan your "scsi" bus (fiber channel is implemented as a scsi device in Linux) to detect the new block devices

    for i in $(ls -1 /sys/class/scsi_host/); do echo '- - -' > /sys/class/scsi_host/$i/scan

  3. Verify you can see the new devices by running multipath -ll

  4. Read, understand, and follow this documentation to complete your setup: http://docs.redhat.com/docs/en-US/Red_Hat_Enterprise_Linux/5/html/DM_Multipath/index.html

I refer you to the documentation for the last step because it's important that you understand how LVM might come into play. You might also want to take advantage of the fact that you can setup an alias for the block device so they are consistent in your environment.

share|improve this answer
    
unfortunately there is no "lpfc_symbolic_name" file under any of /sys/class/scsi_host/host[012] –  Kamil Z Nov 9 '11 at 14:40
    
This information is represented slightly differently in another place. On one of my systems: cat /sys/class/scsi_host/host0/lpfc_symbolic_name Emulex PPN-10:00:00:00:c9:98:37:64 cat /sys/class/scsi_host/host0/device/fc_host\:host0/port_name 0x10000000c9983764 –  JakePaulus Nov 9 '11 at 14:49

You are not seeing your LUNs at all according to your SCSI topology. You need to see the LUNs before you think of multi-pathing. First rescan your HBAs to see if the LUNs are detected and check the switches for zoning and the CX300 for initiator mapping. After the LUNs show up you're good to setup dm accordingly.

Also why aren't you running EMC's PowerPath - that is the EMC multi-pathing drivers? You should get QLogic's SANscreen software on the initiators and the Brocade DCFM server on a central box, makes a few things easier if you are new to FC.

share|improve this answer
    
Yes, but I have 2.6.18 kernel and here i found information that for newer kernels(newer than 2.6.16 as I understood) "You will no longer find the module under /proc/scsi" but under /sys/class/fc_host/. How true is this? I have two subdirectories (host1 and host2) in this directory, and several files there. –  Kamil Z Nov 9 '11 at 14:34

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.