1

My environment comprised AIX 6.1 OS installed over six physical servers and connected via brocade switch to DS8000 SAN storage.

I am aiming to build HACMP cluster on two physical machines in a way that disk could fail-over to each other of two servers. For the reason in DS8000 I connected pool of volumes from one server with WWN FC adapter of the second server.

I ran cfgmgr on the target/second server but The hdisks are not appeared as expected.

How do you recommend to check and handle such problem?

peterh
  • 4,953
  • 13
  • 30
  • 44
Greg P
  • 11
  • 2

1 Answers1

1

I'd check the following in order

  1. Verify HBA adapters on second servers are connected to Brocade switch(s).
  2. Verify WWPN's from 2nd server are logged into Brocade switch ports.
  3. Verify HBA WWPN's on second server are zoned correctly in Brocade switch(s).
  4. Verify WWPN's from 2nd server are visible from DS8000.
  5. Verify LUN's on DS8000 are assigned to the correct WWPN's for the 2nd server.

If all the above is correct then run cfgmgr on AIX host to discover new hdisk devices.

jgruiz
  • 11
  • 1
  • Thanks much pichogve for the prompt and detail answer, I should go to DC and verify there the physical connections between switch storage and the nodes. Meanwhile I tried another available HBA on second node. The HBAs WWPN is member of the Zone and also WWNP connected to volume_group_IDof the target LUN. I run cfgmgr and it reported error message "0514-061 Cannot find a child device". Looks like – Greg P Apr 17 '15 at 17:09
  • It is good for you to check physical connectivity, but there are things you can do just from AIX, for example: on AIX your HBA's will show up as fcsX devices, lets assume one adapter port you are using is fcs0. Then you can use the fcstat fcs0 command to check connectivity. If the fcstat fcs0 command returns an error the port is not connected, if it returns output, look for 'Port Speed (running)' if that shows an actual speed (like 8 GBIT for example) then your port is connected, if not then your cabling is incorrect, or the Brocade port is not enabled. – jgruiz Apr 17 '15 at 19:42
  • I run the command fcstat command and the HBA statistics looks good Class of Service: 3 Port Speed (supported): 4 GBIT Port Speed (running): 4 GBIT Port FC ID: 0xbd0d00 Port Type: Fabric Seconds Since Last Reset: 235568 Transmit Statistics Receive Statistics ------------------- ------------------ Frames: 403 499 Words: 8960 11264 LIP Count: 0 NOS Count: 0 Error Frames: 0 Dumped Frames: 0 Link Failure Count: 1 Loss of Sync Count: 1 Is this enough to assume that the physical connection is OK? – Greg P Apr 18 '15 at 09:56
  • Yes, that is a good start, now look in the Brocade and check the switch port side, and make sure the same WWPN's you see with the fcstat side on AIX are loggin in the switch. – jgruiz Apr 18 '15 at 16:56
  • The port is active and the WWPN is correct in the Brodcade switch. What is strange "lshostconnect -login" command that I run in DS8000 is not showing the HBA statistics (WWNN, WWPN, etc) any more. – Greg P Apr 19 '15 at 07:25
  • I removed from command string of mkhostconnect the volume group parameter and run it again like this "mkhostconnect -dev IBM.2107-75FA120 –wwname WWPN –profile “IBM pSeries – AIX” NAME" as results the HBA WWPN started appearing via lshostconnect but there is still no LUNs/hdisks on the 2nd server. – Greg P Apr 19 '15 at 08:59
  • You still need to assign a volume group to your host, the volume group is what actually allocates LUN's to your host. Take a look at the 'chhostconnect' command. – jgruiz Apr 20 '15 at 12:39