Search This Blog

Monday, June 13, 2011

FC Card Replacement in Superdome Server


 
Primary Details Required:

# fcmsutil /dev/fcd0 | grep "Driver state"
                           Driver state = AWAITING_LINK_UP
# fcmsutil /dev/fcd1 | grep "Driver state"
                           Driver state = AWAITING_LINK_UP
# fcmsutil /dev/fcd2 | grep "Driver state"
                           Driver state = ONLINE
# fcmsutil /dev/fcd3 | grep "Driver state"
                           Driver state = AWAITING_LINK_UP

# fcmsutil /dev/fcd0 | grep "N_Port Port World Wide Name"
            N_Port Port World Wide Name = 0x50060b0000c127e8
# fcmsutil /dev/fcd1 |grep "N_Port Port World Wide Name"
            N_Port Port World Wide Name = 0x50060b0000c127ea
# fcmsutil /dev/fcd2|grep "N_Port Port World Wide Name"
            N_Port Port World Wide Name = 0x50060b00002aa200
# fcmsutil /dev/fcd3|grep "N_Port Port World Wide Name"
            N_Port Port World Wide Name = 0x50060b00002aa202

# ioscan -kfnC fc
Class     I  H/W Path        Driver   S/W State   H/W Type     Description
===========================================================================
fc        0  1/0/10/1/0/4/0  fcd      CLAIMED     INTERFACE    HP AB465-60001 PCI/PCI-X Fibre Channel 2-port 2Gb FC/2-port 1000B-T Combo Adapter (FC Port 1)
                            /dev/fcd0
fc        1  1/0/10/1/0/4/1  fcd      CLAIMED     INTERFACE    HP AB465-60001 PCI/PCI-X Fibre Channel 2-port 2Gb FC/2-port 1000B-T Combo Adapter (FC Port 2)
                            /dev/fcd1
fc        2  1/0/12/1/0/4/0  fcd      CLAIMED     INTERFACE    HP AB465-60001 PCI/PCI-X Fibre Channel 2-port 2Gb FC/2-port 1000B-T Combo Adapter (FC Port 1)
                            /dev/fcd2
fc        3  1/0/12/1/0/4/1  fcd      CLAIMED     INTERFACE    HP AB465-60001 PCI/PCI-X Fibre Channel 2-port 2Gb FC/2-port 1000B-T Combo Adapter (FC Port 2)
                            /dev/fcd3

# olrad -q
                                                        Driver(s)
                                                        Capable
Slot     Path            Bus   Max  Spd  Pwr  Occu Susp OLAR  OLD  Max   Mode
                         Num   Spd                                 Mode
0-0-3-8  1/0/12/1        352   133  133  On   Yes  No   Yes   N/A  PCI-X PCI-X
0-0-3-9  1/0/10/1        336   133  133  On   Yes  No   Yes   N/A  PCI-X PCI-X
0-0-3-10 1/0/9/1         328   133  133  On   Yes  No   Yes   N/A  PCI-X PCI-X
0-0-3-11 1/0/8/1         320   133  133  On   Yes  No   Yes   N/A  PCI-X PCI-X

Cabinet#-Bay#-Chassis#-Slot#

# olrad -h 0-0-3-9 
1/0/10/1/0/4/0-------This one is faulty
1/0/10/1/0/4/1
1/0/10/1/0/6/0
1/0/10/1/0/6/1

Card location is 0-0-3-9
Card is Located at Cabinet 0, Bay 0 (Front), Chasis 3(Right), Slot Number 9

# lltstat -nvv|more
LLT node information:
    Node                 State    Link  Status  Address
   * 0 testnodep          OPEN   
                                  lan4   UP      00:17:A4:A4:25:4A
                                  lan6   UP      00:17:A4:A4:E4:DA
                                  lan900 UP      00:1E:0B:83:01:1A
     1 testnodec          OPEN   
                                  lan4   UP      00:17:A4:A4:25:22
                                  lan6   UP      00:17:A4:A4:65:92
                                  lan900 UP      00:1E:0B:83:31:70

# lanscan -qv|grep -i lan4
1/0/10/1/0/6/0 0x0017A4A4254A 4   UP    lan4 snap4     5   ETHER     Yes     119

# lanscan -qv|grep -i lan6
1/0/12/1/0/6/0 0x0017A4A4E4DA 6   UP    lan6 snap6     7   ETHER     Yes     119

# /opt/VRTSvcs/bin/hastatus -summ

-- SYSTEM STATE
-- System               State                Frozen             

A  testnodep             RUNNING              0                    
A  testnodec             RUNNING              0                   

-- GROUP STATE
-- Group           System               Probed     AutoDisabled    State         

B  IMRMLESBPRD     testnodep             Y          N               ONLINE        
B  IMRMLESBPRD     testnodec             Y          N               OFFLINE       

Implementation Summary:

Note: This change can be done on-line with no impact to service

1. Log call with HP to replace card
2. Offline / Power off the FC card
3. Replace FC Card
4. Power on / Online FC Card
5. Reconfigure Network APA
6. Storage Management to re-zone to new WWN


Make a note of the existing WWN relating to the faulty card. This will help to ensure that the correct zoning and LUN security can be carried out by storage Management.

# fcmsutil /dev/fcd0 | grep "N_Port Port World Wide Name"
N_Port Port World Wide Name = 0x50060b0000c127e8


When the FC card is ready to be replaced offline the card as follows:

Note: If Oracle is not running on testnodec or it is not using the ODM, the ODM can be stopped without affecting the service.

Stop the ODM

# /sbin/init.d/odm stop
# umount /dev/odm

# lvlnboot -v
Boot Definitions for Volume Group /dev/vg00:
Physical Volumes belonging in Root Volume Group:
        /dev/dsk/c13t0d0 (1/0/12/1/0/4/0.215.40.0.0.0.0) -- Boot Disk
        /dev/dsk/c13t0d1 (1/0/12/1/0/4/0.215.40.0.0.0.1)
        /dev/dsk/c17t0d0 (1/0/10/1/0/4/0.199.40.0.0.0.0)
        /dev/dsk/c17t0d1 (1/0/10/1/0/4/0.199.40.0.0.0.1)
Boot: lvol1     on:     /dev/dsk/c13t0d0
                        /dev/dsk/c17t0d0
Root: lvol3     on:     /dev/dsk/c13t0d0
                        /dev/dsk/c17t0d0
Swap: lvol2     on:     /dev/dsk/c13t0d0
                        /dev/dsk/c17t0d0
Dump: lv_dump   on:     /dev/dsk/c13t0d0, 0


Before HBA replaced

# pvchange -a n /dev/dsk/c17t0d0
# pvchange -a n /dev/dsk/c17t0d1

1.        Display Current PCI Card:

# olrad –q

  1. Run a CRA on the slot. A CRA is a Critical Resources Analysis which means the system will be checked to make sure that powering off the card in the specified slot will not lose data, network access and whole load of the other checks:

# olrad –C 0-0-3-9


  1. All I/O on the card are duplicated on another card so the card is ok to be powered off.

# olrad –f –r 0-0-3-9 

The card is now safe to removed and replaced by the HP engineer.

  1. When replaced switch the power back on the card and have the system initialize the drivers:

# olrad –R 0-0-3-9 

  1. Start /Mount the odm

# /sbin/init.d/odm start 
# mount /dev/odm

  1. Set the lan4 interface to 100fd:

Syntax: lanadmin –X 100fd <lan device>
#lanadmin –X 100fd 2


  1. Storage Management needs to re-zone and apply LUN security to the New WWN.
Inform Storage Management that the zoning and LUN security for the old WWN can be removed. Old WWN is 0x50060b0000c127e8

Get the new WWN

Syntax: fcmsutil <HBA device file>|grep "N_Port Port World Wide Name"
#fcmsutil /dev/fcd0 |grep “N_Port Port World Wide Name"


Inform Storage Management of new WWN and they will re-zone and apply LUN security

  1. Check the card has logged into the SAN fabric

Syntax: fcmsutil <HBA device file>|grep "Driver state"
# fcmsutil /dev/fcd0|grep "Driver state"

Outputs of above should be ONLINE

Re-Enable the path to the disks on new HBA
Check disk are CLAIMED on path 1/0/10/1/0/4/0

# ioscan –fnH 1/0/10/1/0/4/0

  1. If outputs from above shows that disk are in CLAIMED STATE bring the disk under vxvm control.
Veritas volume:

# vxdctl enable

LVM Volume:
After HBA replqced

# pvchange -a y /dev/dsk/c17t0d0
# pvchange -a y /dev/dsk/c17t0d1

 Than as yo have already specified reactivate vg00

# vgchange -a y /dev/vg00
# lvlnboot -R /dev/vg00

No comments:

Post a Comment