How to migrate GPFS from 3.5 to 4.2 in AIX?

How to Migrate GPFS version from 3.5 to 4.2 in AIX?

How to upgrade GPFS version from 3.5 to 4.2 in AIX?

The GPFS is a high-performance clustered file system that can be deployed in shared-disk. GPFS provides concurrent high-speed file access to the applications executing on multiple nodes of clusters which can be used with AIX5L, Linux, and Windows platforms.

Note: We didn't see anywhere for the specific step by step document for upgrading the gpfs from 3.5 to 4.2. Myself and my Boss spent much time to prepare these steps for upgrade the gpfs. Hope it is helpful.

1.   Ground work:

mmgetstate                  To displays the state of the GPFS daemon on one or more nodes

lslpp -l gpfs                  To list the current gpfs version details

mmdf                           To check the gpfs file system size

mmlsfs all                      To check all GPFS file systems

mmlscluster                 To displays GPFS cluster configuration information

mmlsconfig                   To displays the configuration data for a GPFS cluster

mmlsmgr                      To displays the file system manager node

mmlsmgr  -c                 To view the GPFS manager

mmlsnsd                       To list NSD disks

mmlsnsd -M                  To view the detailed NSD disks

mmlsdisk -d                  To view the disk information

mmlslicense                 To view the GPFS license

mmlsmgr                      To check the cluster Manager and File Manager

lspv                              To get the disk details of the server.

installp -s                     To check any filesets are in applied state

installp -c all                To commit all the filesets if anything is on the applied state.

prtconf                         To get the basic information about the server.

2.   Stop GPFS cluster and uninstall the current version

Stop all user activity in the file systems.

#mmexportfs all –o /var/tmp/exporDataFile     To export the current cluster configuration

#mmshutdown –a                                  To Stop GPFS on all nodes in the cluster.

#lslpp –l|grep –I gpfs                             To identify the current level of the gpfs 3.5

#Installp –u gpfs                                    To uninstall the gpfs on all the nodes.

#lslpp –l|grep –I gpfs                             To confirm the gpfs is completely uninstall on servers.

#shutdown –Fr                                     For each node, initiate the reboot of the servers.

3.   Upgrade to GPFS 4.2 (Spectrum Scale)

#lslpp –l|grep –I gpfs               To identify the current level of the gpfs

#cd /var/tmp/gpfs_4.2             To go the particular dir where we keep the 4.2 filesets.

#smitty update_all                   Do the preview first and commit next, perform same on all the nodes.

#lslpp –l|grep –I gpfs                To confirm the gpfs 4.2 has been installed on all the nodes.

#mmgetstate –a                       To check the gpfs cluster status after 4.2 installation

#mmstartup –a                         To start the gpfs cluster

#mmgetstate –a                       To check the gpfs cluster status after 4.2 installation

#df –gt  /data/test                    To confirm the gpfs filesytem mounted automatically (if not, follow the below)

#mmimportfs all –I /var/tmp/exportDataFile   To import the gpfs filesytem configuration details.

#mmmount /dev/testlv –a         To manually mount the gpfs filesytem

#df –gt /data/test                     To confirm the gpfs filesytem mounted all the nodes.

#shutdown –Fr                         To reboot all the nodes.

4.   Validation

#mmgetstate –a                         To confirm the cluster service is active on all the nodes.

#df –gt/data/test                       To confirm the cluster filesystem mounted on all the nodes.

#lslpp –l |grep –I gpfs              To confirm the latest level of gpfs is in commited state.

Ask application team to peform the technical checks out of the servers and confirm all is OK at their end.

5. Post migration tasks to the new level of spectrum scale

#mmlscconfig                                              At the moment, the filesytem configuration still shows as 3.5

#mmchconfig release=LATEST                     To migrate the cluster configuration data and enable new functionality throughout the cluster

#mmchconfig release=LATEST --accept-empty-cipherlist-security  To execute this command only if the above mmchconfig command failed.

#mmlslicense -L                                          To get GPFS license is for each of the nodes in the cluster

#mmchlicense server -N NodeList               To assign a GPFS client license to the nodes that require it

#mmchfs <FileSystem> -V compat              To enable backward-compatible format changes

(or)

#mmchfs FileSystem -V full                        To migrate all file systems to the latest metadata format changes

Note:

If you issue mmchfs -V compat, only changes that are backward compatible with GPFS 3.5 will be enabled. Nodes in remote clusters that are running GPFS 3.5 will still be able to mount the file system.

If you issue mmchfs -V full, all new functions that require different on-disk data structures will be enabled. Nodes in remote clusters running an older GPFS version will no longer be able to mount the file system

6. Backout steps for the gpfs from 4.2 to 3.5

#mmexportfs all –o /tmp/exporDataFile  To export the current cluster configuration

#lslpp –l|grep –I gpfs                           To check the current level of the gpfs

Installp –u gpfs                                    To uninstall the current 4.2 gpfs on all the nodes.

#lslpp –l|grep –I gpfs                           To confirm the gpfs 4.2 is completely uninstall on servers.

#shutdown –Fr                                    To reboot all the nodes.

#cd /var/tmp/gpfs_3.5                        To go the particular dir where we keep the 3.5 filesets.

#smitty update_all                              Do the preview first and commit next, perform the same on all the nodes.

#lslpp –l|grep –I gpfs                             To confirm the gpfs 3.5 has been installed on all the nodes.

#mmgetstate –a                                    To check the gpfs cluster status after 3.5 installation

#mmstartup –a                                      To start the gpfs cluster

#mmgetstate –a                                    To check the gpfs cluster status after 3.5 installation

#df –gt  /data/test                                To confirm the gpfs filesytem mounted automatically (if not, follow the below)

#mmimportfs all –I /tmp/exportDataFile   To import the gpfs filesytem configuration details.

#mmmount /dev/testlv –a                     To manually mount the gpfs filesytem

#df –gt /data/test                                  To confirm the gpfs filesytem mounted all the nodes.

#shutdown –Fr                                      To reboot all the nodes.

Ask application team to perform the technical checks out and get confirmation that all is OK 

How to upgrade HACMP version in aix?

How to upgrade HACMP - Cluster version in aix?

1. Ground work:

#lslpp -l|grep -i cluster.es.server.rte     --> To check the current version of the cluster

#./clRGinfo                            --> Check if the RG is online status.

#lssrc -ls clstrmgrES               --> check if the cluster manager is stable on both nodes

#smitty hacmp ->

    Extended configuration

      snapshot configuration

        Create a snapshot of the cluster configuration  (Mandatory -snap name and description)

             output: .info and .odm file created (Required this for restoration)

#/var/hacmp/clverify/clverify.log   --> check clverify has no errors

#emgr -l                                          --> Check if the emergency fixes there on the server

#emgr -rL                                          --> To remove the emergency fix.

Check Download the hacmp upgrade filesets from fixcentral and save it in rootvg directory.

2. Implementation steps

Stop the cluster services on both the nodes.

#lssrc -ls clstrmgrES                    --> check if the cluster manager is st_stable on both nodes

#smitty hacmp

     system management (c-spoc)

         manage hacmp services

             stop cluster services  (now, select node/noes and bring resource groups offline)

#lssrc -ls clstrmgrES                    --> Now the cluster manager demon should be ST_INIT

#lslpp -l|grep -i cluster.es.server.rte   --> To check the current version of hacmp

#cd <package_located directory>

#smitty update_all        -- > Do the preview first and commit next, once done follow below

#lslpp -l|grep -i cluster.es.server.rte    --> To confirm the version has been upgraded.

Reboot the node and perform the upgrade on the other node.

#smitty hacmp

     system management (c-spoc)

         manage hacmp services

             start cluster services  (now, select node/node names, automatically)

#/tmp/hacmp.out                      --> To check any error and warnings

3. Validation

#lslpp -l|grep -i cluster.es.server.rte    --> To confirm the version has been upgraded.

#lppchk –v                                           --> No output should displayed, only the prompt

#./clRGinfo                                         --> To check the RG is online as before.

#lssrc -ls clstrmgrES                            --> To check if the cluster manager is stable (ST_Stable)

How to update vio server in aix?

How to upgrade vio server in aix?

Pre-requisite:

Make sure to check the existing vio version $ioslevel

Collect all the connected lpars to the VIO server

Take the mksysb backup of lpars and backupios, viosbr backup of vio servers.

Take note of the important output configuration files and lsmap -all output before the upgrade start.

To get the vio upgrade package from the below site or from your nim server if you got already.

 http://www14.software.ibm.com/webapp/set2/sas/f/vios/download/home.html

  

Implementation steps:

Copy the vio upgrade package to the /home/padmin of the VIO servers.

The upgrade package has to be placed in /home/padmin

Login as padmin mode.

$updateios -commit    --> To Commit previous updates by running the updateios command.

$updateios –accept –install –dev /home/padmin/<vio package>   --> To upgrade the vio version.

$shutdown -restart   -->Reboot the vio server to complete the update.

$ioslevel     --> Check the latest vio level has been achieved.  

Important notes:

When upgrading to VIOS 2.2.0.10-FP-24

 Fix Pack 24 enables Role Based Access Control (RBAC) roles and authorizations for thepadmin user on the VIOS. Roles and authorizations are enabled immediately. If you upgrade to VIOS 2.2.0.10-FP-24 via the updateios command, the only two commands that you can run before the VIOS is rebooted are the shutdown and swrole commands.

$ swrole - PAdmin 

$ shutdown -restart

$ license -accept

    

How to upgrade the adapter's firmware level in aix?

Firmware upgrade for adapter in aix?

1. List the fibre channel adapters installed in the system 

lsdev -C | grep fcsX  where fcsX is the adapter you on which you want to install the microcode


2. Determine the current microcode level on the adapter 

lsmcode -d fcsX 


3. Download the Microcode RPM Package and put it into the temporary directory and Unpacking it using the below command.

rpm -ivh --ignoreos <rpm_package>   (If the microcode package unpacks successfully, the microcode file will be added to the /etc/micrococde directory)


4. Update the adapter's microcode level

"diag -d fcsX -T download" 


5. Confirm the current microcode level.

lsmcode -d fcsX

How to upgrade TL and SP in AIX?

 How to upgrade TL and SP in AIX?

1. Ground work:

oslevel -s                                 --> find out the current os level of the system.

instfix –i|grep –i ml                --> check currently installed ML levels are consistent

lppchk –vm3                           --> check currently installed filesets are consistent

df -g                                         --> check the system has the enough free space.

bootlist -m normal -o             --> check the blv has been created on the rootvg disk
emgr -l                                     --> check the installed ifixes/emergency fixes.
emgr -r -L <ifix label>           --> To removing the ifixes / emergency fixes.

Ensure the latest "mksysb" has been taken.

Download the TL/SP package from IBM fix central and put it to nim server

2. Create the alt_disk (Consider we have hdisk0 & hdisk1 on the rootvg)

sysdumpdev –s /dev/sysdumpnull    --> If secondary dump device resides on hdisk1.

rmlv dump_lv                                    --> Remove the dump lv

unmirrorvg rootvg hdisk1                --> unmirrror (Have to use the hdisk1 for alt_disk)

chpv –c hdisk1                                   --> clear the boot image on the hard disk1.

reducevg rootvg hdisk1                    --> remove the hdisk1 from the rootvg disk.

alt_disk_copy -d hdisk1                    --> Take the alt disk clone on hdisk1

lspv                                                     --> confirm the "altinst_rootvg"  created on hdisk1

3. Perform the TL upgrade

smitty commit                                     --> commit all the old applied filesets

installp -s                                            --> check if any os filesets in applied mode

mount <nim_server>:<package holding directory> /mnt  --> mount directory which hold the TL pkg. 

cd /mnt

smitty update_all                               -- > Do the preview first and commit next, once done follow below

oslevel -s                                            --> Check the new TL level.

lppchk -v                                            --> No output should displayed, only the prompt

bootlist -m normal hdisk0             --> Change the bootlist to hdisk0 - remember alt_dsk reside on hdisk1

shutdown -Fr                                      --> for fast reboot.

4. Validation

oslevel –s                                           --> check the new TL level

instfix –i|grep  ML                            --> confirm the new TL level are consistent.

lppchk –v                                           --> No output should displayed, only the prompt

5. Remove the alt_disk and re-mirror

alt_disk_install -X                            --> Remove the alt_disk

extendvg -f rootvg hdisk1                 --> extend the hdisk1 to the rootvg

mirrorvg rootvg hdisk1                    --> To mirror with the hdisk1

bosboot –ad /dev/hdisk1                   --> Create the boot image

bootlist –m normal  hdisk0 hdisk1  --> Set the boot sequence

bootlist –m normal –o                       --> check the boot sequence order

mklv –y <lvname> -t sysdump rootvg <num of LP’s> hdisk1  --> Create the dump lv on hdisk1 if you have removed during the alt disk clone

sysdumpdev –s /dev/<lv name>        --> Create the secondary dump device on hdisk1

sysdumpdev –l                                   --> confirm the dumplv

How to upgrade the firmware level through HMC?

Steps to upgrade the firmware level through HMC


Check the existing firmware level of the servers (lpars) in the managed systems.

1. Login to the HMC with right privileges.

2. Click the managed server which you want to upgrade the firmware level.

3. Click the lpar and do shutdown, few minutes later, you can see the status as - "not activated"

4. poweroff and poweron the managed systems (To make it "STANDBY" mode)

5. Click the "change LIC for the current release "  (You can see this under the upgrade menu in the bottom of the HMC window0

start change licensed internal code wizard (press 0k)

        select ftp site (press ok)

                check the directory of the release (which is hold the package) (Also make sure the ftp site and user id and give the  passwd to proceed)  (press ok)

                        Press OK if you found no error in the next screen

                                press "Next"

                                        press "next"

                                                Accept the license (press "accept")

                                                        select the managed systems name on the check box and press finish to start the upgradation.

                                                                press "ok"  

                                                                        you can see the dialouge box of upgradation. (on that it mentioned 

In here the process automatically poweroff the managed system and prepare to install once the installation completed, the managed systems will be in the power on state.

Start the lpars which is hold by the managed systems. (That we already shutdown when we start this firmware upgradation see number:3)

Login the server and check it out the firmware level using "#lsmcode -c".