Why should I migrate?
Before I discuss how to migrate to a newer version of the IBM AIX® operating system, let's review why you should consider migrating at all. One of the most important reasons to migrate is support. If you are running an older version of AIX, such as 5.3 or earlier, you should already be aware that these versions are no longer supported by IBM. If problems arise on an older version of AIX you are most likely going to be on your own. IBM support will no longer be able to help you.
AIX 5.3 officially went out of support on April 30th 2012. IBM is offering extended support for a limited time for a fee. This will help some customers in the interim while they migrate their systems to AIX 7.1 or 6.1. If you have IBM POWER7® hardware, you might want to consider AIX 5.2 and 5.3 Workload Partitions (WPARs). This will allow you to continue running your legacy applications on AIX 5.2 or 5.3, however they will run within an AIX 7.1 WPAR. These special systems are known as versioned WPARs and are only supported on POWER7 hardware. You can find more information on versioned WPARs.
Of course there are other pressing reasons to migrate as well. Newer releases of AIX include a multitude of new enhancements, improvements, features and performance boosts. I encourage you to read the >AIX 7.1 and 6.1 Differences Guides (IBM Redbooks® publications) to find out more – see the Resources section .
Migrating to AIX 7.1 with nimadm
I've discussed using nimadm to migrate to AIX 6.1 in the past. In this article I'll briefly cover that same process but this time migrating to AIX 7.1. Admittedly the steps are almost identical. So I'll refer you to my 2010 article as a starting point for migrating to 7.1. In this article I'll provide a brief guide to migrating both AIX 5.3 and 6.1 systems to 7.1. I'll also offer some general advice and tips.
The nimadm utility offers several advantages over a conventional migration. For example, a system administrator can use nimadm to create a copy of a NIM client's rootvg (on a spare disk on the client, similar to a standard alternate disk installation alt_disk_install) and migrate the disk to a newer version or release of AIX. All of this can be done without disruption to the client (there is no outage required to perform the migration). After the migration is finished, the only downtime required will be a scheduled reboot of the system.
Another advantage is that the actual migration process occurs on the NIM master, taking the load off the client client logical partition (LPAR). This reduces the processing overhead on the LPAR and minimizes the performance impact to the running applications.
For customers with a large number of AIX systems, it is also important to note that the nimadm tool supports migrating several clients at once.
Just as I did in my previous article I'll assume you already have a NIM master in your environment. And I'm going to assume that your NIM master is already running AIX 7.1 with the latest latest technology level (TL) and and service pack (SP) applied. If not, I recommend you refer to my previous article and the associated Resources section.
My NIM master is running AIX 7.1 TL1 SP4.
I created new lpp_source and SPOT NIM resources for AIX 7.1 TL1 SP4.
# lsnim -t lpp_source lpp_sourceaix710104 resources lpp_source # lsnim -t spot spotaix710104 resources spot # lsnim -l lpp_sourceaix710104 lpp_sourceaixaix710104: class = resources type = lpp_source arch = power Rstate = ready for use prev_state = unavailable for use location = /export/lpp_source/lpp_sourceaix710104 simages = yes alloc_count = 0 server = master # lsnim -l spotaix710104 spotaix7101014: class = resources type = spot plat_defined = chrp arch = power Rstate = ready for use prev_state = verification is being performed location = /export/spot/spotaix7101014/usr version = 7 release = 1 mod = 1 oslevel_r = 7100-01 alloc_count = 2 server = master if_supported = chrp.64 ent Rstate_result = success
To create these resources, I downloaded AIX 7.1 ISO images from the IBM Entitled Software support website. I placed these images into a temporary directory and then used the loopmount command to temporarily mount them.
# ls -ltr total 11301248 drwxr-xr-x 2 root system 256 May 14 15:47 lost+found -rw-r--r-- 1 root system 3361046528 May 14 16:22 NIMAIX71DVD1.iso -rw-r--r-- 1 root system 2425192448 May 14 16:23 NIMAIX71DVD2.iso # loopmount -i NIMAIX71DVD1.iso -o "-V cdrfs -o ro" -m /mnt/dvd1 # loopmount -i NIMAIX71DVD2.iso -o "-V cdrfs -o ro" -m /mnt/dvd2 # df | grep loop /dev/loop0 6563484 0 100% 1640871 100% /mnt/dvd1 /dev/loop1 4735648 0 100% 1183912 100% /mnt/dvd2 # ls -ltr /mnt/dvd1 /mnt/dvd2 /mnt/dvd1: total 84 drwxr-xr-x 3 4000 4000 2048 Sep 03 2010 ppc -rw-r--r-- 1 4000 4000 819 Sep 03 2010 README.aix drwxr-xr-x 2 4000 4000 2048 Sep 03 2010 7100-00 drwxr-xr-x 3 4000 4000 2048 Sep 03 2010 root -rw-r--r-- 1 4000 4000 15081 Sep 03 2010 image.data -rw-r--r-- 1 4000 4000 6252 Sep 03 2010 bosinst.data -rw-r--r-- 1 4000 4000 16 Sep 03 2010 OSLEVEL drwxrwxr-x 4 4000 4000 2048 Sep 03 2010 RPMS drwxr-xr-x 11 4000 4000 2048 Sep 03 2010 usr drwxr-xr-x 4 4000 4000 2048 Sep 03 2010 installp -rw-rw-r-- 1 4000 4000 42 Sep 03 2010 .Version /mnt/dvd2: total 16 drwxr-xr-x 3 4000 4000 2048 Sep 03 2010 ismp drwxrwxr-x 3 4000 4000 2048 Sep 03 2010 usr drwxrwxr-x 4 4000 4000 2048 Sep 03 2010 installp -rw-rw-r-- 1 4000 4000 42 Sep 03 2010 .Version
After they are mounted, I used smittybffcreate to copy the contents of these images to my new AIX 7.1 lpp_source directory create (/export/lpp_source/lpp_sourceaix710104).
Copy Software to Hard Disk for Future Installation
Type or select values in entry fields. Press Enter AFTER making all desired changes. [Entry Fields] * INPUT device / directory for software /mnt/dvd1 * SOFTWARE package to copy [all] + * DIRECTORY for storing software package [/export/lpp_source/lpp_sourceaixaix710104] DIRECTORY for temporary storage during copying [/tmp] EXTEND file systems if space needed? yes + Process multiple volumes? yes +
After the base filesets had been copied over, I defined the new lpp_source and SPOT within NIM. Then I downloaded the latest TL and SP for AIX 7.1 and updated the lpp_source and SPOT.
My NIM clients are running AIX 5.3 and AIX 6.1. We will migrate both of these systems to AIX 7.1 using nimadm. We will review the migration process for each system and check that the systems are appropriately tuned after the migration.
root@lparlparaix53[/] > oslevel -s 5300-12-04-1119 root@lparlparaix61[/] > oslevel -s 6100-06-04-1112
I have NIM client definitions for both systems already in place on my NIM master.
# lsnim -t standalone lparlparaix53 machines standalone lparlparaix61 machines standalone # lsnim -l lparaix53 lparaix53: class = machines type = standalone connect = nimsh platform = chrp netboot_kernel = 64 if1 = 172_29_154 lparaix53 0 cable_type1 = N/A Cstate = ready for a NIM operation prev_state = not running Mstate = currently running cpuid = 00CDB5114C00 Cstate_result = reset # lsnim -l lparaix61 lparaix61: class = machines type = standalone connect = nimsh platform = chrp netboot_kernel = 64 if1 = 172_29_154 lparaix61 0 cable_type1 = N/A Cstate = ready for a NIM operation prev_state = not running Mstate = currently running cpuid = 00C8E4244C00 Cstate_result = reset
Each NIM client has a spare disk available that we will use for the alternate disk migration.
root@lparaix53[/] > lspv hdisk4 00f6050a2cd79ef8 rootvg active hdisk5 00cdb511757d999e None root@lparaix61[/] > lspv hdisk4 00f6050a2cd79ef8 rootvg active hdisk5 00c8e42485fabfd7 None
The NIM Master is configured with only three volume groups, rootvg, nimvg and nimadmvg. The nimadmvg volume group will be used as temporary client data cache location for the 7.1 migrations. The volume group is currently empty.
# lsvg rootvg nimvg nimadmvg # lspv | grep adm hdisk4 00c342c6395ff736 nimadmvg active # lsvg -l nimadmvg nimadmvg: LV NAME TYPE LPs PPs PVs LV STATE MOUNT POINT
Currently, nimadm requires rsh access to the NIM clients in order to function. Therefore we ensure that the NIM Master has rsh access to each of the clients and that NIM can communicate with each client successfully.
# rsh lparaix53 date Fri May 18 14:06:07 EETDT 2011 # rsh lparaix61 date Sat May 18 14:06:40 EETDT 2012 # nim -o lslpp lparaix53 | grep -w "bos.rte " bos.rte 22.214.171.124 COMMITTED Base Operating System Runtime bos.rte 126.96.36.199 COMMITTED Base Operating System Runtime # nim -o lslpp lparaix61 | grep -w "bos.rte " bos.rte 188.8.131.52 COMMITTED Base Operating System Runtime bos.rte 184.108.40.206 COMMITTED Base Operating System Runtime bos.rte 220.127.116.11 COMMITTED Base Operating System Runtime bos.rte 18.104.22.168 COMMITTED Base Operating System Runtime bos.rte 22.214.171.124 COMMITTED Base Operating System Runtime bos.rte 126.96.36.199 COMMITTED Base Operating System Runtime bos.rte 188.8.131.52 COMMITTED Base Operating System Runtime bos.rte 184.108.40.206 COMMITTED Base Operating System Runtime
We will now initiate a migration for both systems. Starting with the 5.3 system, we run the following nimadm command on the NIM master to start the alternate disk-migration process.
NIMADM operation for AIX 5.3 system:
# nimadm -j nimvadmg -c lparaix53 -s spotaix710104 –l lpp_sourceaix710104 -d "hdisk5" –Y Initializing the NIM master. Initializing NIM client lparaix53. Verifying alt_disk_migration eligibility. Initializing log: /var/adm/ras/alt_mig/lparaix53_alt_mig.log Starting Alternate Disk Migration. +-----------------------------------------------------------------------------+ Executing nimadm phase 1. +-----------------------------------------------------------------------------+ Cloning altinst_rootvg on client, Phase 1. Client alt_disk_install command: alt_disk_copy -j -M 7.1 -P1 -d "hdisk5" Calling mkszfile to create new /image.data file. Checking disk sizes. Creating cloned rootvg volume group and associated logical volumes. Creating logical volume alt_hd5 Creating logical volume alt_hd6 Creating logical volume alt_hd8 ...ETC...
In a new secure shell (SSH) session on the NIM master we initiate a second nimadm operation to migrate our AIX 6.1 NIM client.
NIMADM operation for AIX 6.1 system:
# nimadm -j nimadmvg -c lparaix61 -s spotaix710104 –l lpp_sourceaix710104 -d "hdisk5" –Y Initializing the NIM master. Initializing NIM client lparaix61. Verifying alt_disk_migration eligibility. Initializing log: /var/adm/ras/alt_mig/lparaix61_alt_mig.log Starting Alternate Disk Migration. +-----------------------------------------------------------------------------+ Executing nimadm phase 1. +-----------------------------------------------------------------------------+ Cloning altinst_rootvg on client, Phase 1. Client alt_disk_install command: alt_disk_copy -j -M 7.1 -P1 -d "hdisk5" Calling mkszfile to create new /image.data file. Checking disk sizes. LOGICAL_VOLUME= hd11admin FS_LV= /dev/hd11admin Creating cloned rootvg volume group and associated logical volumes. Creating logical volume alt_hd5 Creating logical volume alt_hd6 Creating logical volume alt_hd8 ...ETC...
At this point the nimadm process is copying data from the clients, to the cache file systems on the NIM master, and performing the migration on the master itself. After the migration process for a client is complete, the data is copied back to the client's alternate rootvg disk. If you are interested in learning more about each phase of the nimadm process, then again, I refer to my original nimadm article on the IBM developerWorks® website.
We observe that both clients now have a new volume group named altinst_rootvg. This volume group contains a copy of the original rootvg, now migrated to AIX 7.1.
root@lparaix53[/tmp] > lspv hdisk4 00f6050a2cd79ef8 rootvg active hdisk5 00cdb511757d999e altinst_rootvg active root@lparaix61[/] > lspv hdisk4 00f6050a2cd79ef8 rootvg active hdisk5 00c8e42485fabfd7 altinst_rootvg active
As the migrated data is being copied from the NIM master to the client, we observe that the alternate rootvg file systems are temporarily mounted on each client to receive the data.
On the NIM Master we discover temporary cache file system mounts for each of the NIM clients. These cache file systems are housed in the nimadmvg volume group.
# df -g Filesystem GB blocks Free %Used Iused %Iused Mounted on /dev/hd4 0.38 0.16 58% 11761 22% / /dev/hd2 6.00 1.95 68% 93348 17% /usr /dev/hd9var 2.38 2.02 15% 7620 2% /var /dev/hd3 0.50 0.49 3% 187 1% /tmp /dev/hd1 0.12 0.10 24% 18 1% /home /dev/hd11admin 0.12 0.12 1% 7 1% /admin /proc - - - - - /proc /dev/hd10opt 2.62 2.01 24% 12726 3% /opt /dev/livedump 0.25 0.25 1% 4 1% /var/adm/ras/livedump /dev/cglv 25.00 19.61 22% 6 1% /cg /dev/lppsrclv 25.00 13.99 45% 7269 1% /lppsrc /dev/spotlv 25.25 23.96 6% 29122 1% /spot /dev/loop0 3.13 0.00 100% 1640871 100% /mnt/dvd1 /dev/loop1 2.26 0.00 100% 1183912 100% /mnt/dvd2 /dev/lv00 0.25 0.16 37% 7441 12% /lparaix53_alt/alt_inst /dev/lv01 0.25 0.24 4% 18 1% /lparaix53_alt/alt_inst/admin /dev/lv02 0.25 0.24 4% 25 1% /lparaix53_alt/alt_inst/home /dev/lv03 0.75 0.29 62% 6718 4% /lparaix53_alt/alt_inst/opt /dev/lv04 0.25 0.20 21% 225 1% /lparaix53_alt/alt_inst/tmp /dev/lv05 3.50 0.73 80% 80350 9% /lparaix53_alt/alt_inst/usr /dev/lv06 0.25 0.15 42% 7792 12% /lparaix53_alt/alt_inst/var /dev/lv07 0.25 0.03 89% 13533 21% /lparaix61_alt/alt_inst /dev/lv08 0.25 0.24 4% 17 1% /lparaix61_alt/alt_inst/admin /dev/lv09 0.25 0.24 4% 23 1% /lparaix61_alt/alt_inst/home /dev/lv10 0.75 0.61 19% 6409 4% /lparaix61_alt/alt_inst/opt /dev/lv11 0.25 0.24 6% 107 1% /lparaix61_alt/alt_inst/tmp /dev/lv12 4.25 0.26 94% 104878 10% /lparaix61_alt/alt_inst/usr /dev/lv13 0.25 0.12 54% 9917 16% /lparaix61_alt/alt_inst/var /usr/lib 6.00 1.95 68% 93348 17% /lparaix53_alt/alt_inst/usr/ lib/alt_mig/usr/lib /usr/ccs/lib 6.00 1.95 68% 93348 17% /lparaix53_alt/alt_inst/usr/ lib/alt_mig/usr/ccs/lib /lparaix53_alt/alt_inst 0.25 0.16 37% 7441 12% /lparaix53_alt/alt_inst/usr/ lpp/bos/inst_root /lparaix53_alt/alt_inst/var 0.25 0.15 42% 7792 12% /lparaix53_alt/alt_inst/usr/ lpp/bos/inst_root/var /lparaix53_alt/alt_inst/tmp 0.25 0.20 21% 225 1% /lparaix53_alt/alt_inst/usr/ lpp/bos/inst_root/tmp /lparaix53_alt/alt_inst/home 0.25 0.24 4% 25 1% /lparaix53_alt/alt_inst/usr/ lpp/bos/inst_root/home /lparaix53_alt/alt_inst/admin 0.25 0.24 4% 18 1% /lparaix53_alt/alt_inst/usr/ lpp/bos/inst_root/admin # lsvg -l nimadmvg nimadmvg: LV NAME TYPE LPs PPs PVs LV STATE MOUNT POINT loglv00 jfs2log 1 1 1 open/syncd N/A lv00 jfs 1 1 1 open/syncd /lparaix53_alt/alt_inst lv01 jfs 1 1 1 open/syncd /lparaix53_alt/alt_inst/admin lv02 jfs 1 1 1 open/syncd /lparaix53_alt/alt_inst/home lv03 jfs 3 3 1 open/syncd /lparaix53_alt/alt_inst/opt lv04 jfs 1 1 1 open/syncd /lparaix53_alt/alt_inst/tmp lv05 jfs 14 14 1 open/syncd /lparaix53_alt/alt_inst/usr lv06 jfs 1 1 1 open/syncd /lparaix53_alt/alt_inst/var lv07 jfs 1 1 1 open/syncd /lparaix61_alt/alt_inst lv08 jfs 1 1 1 open/syncd /lparaix61_alt/alt_inst/admin lv09 jfs 1 1 1 open/syncd /lparaix61_alt/alt_inst/home lv10 jfs 3 3 1 open/syncd /lparaix61_alt/alt_inst/opt lv11 jfs 1 1 1 open/syncd /lparaix61_alt/alt_inst/tmp lv12 jfs 17 17 1 open/syncd /lparaix61_alt/alt_inst/usr lv13 jfs 1 1 1 open/syncd /lparaix61_alt/alt_inst/var
Each of my migrations took around 60 minutes to complete. By using nimadm, that was an hour of downtime I was able to avoid.
We can review the associated nimadm log files for each client to verify the migration process was successful. You can monitor the migration by tailing each of the log files on the NIM master.
# cd /var/adm/ras/alt_mig # ls –ltr *.log total 7136 -rw-r--r-- 1 root system 420858 May 18 14:33 lparaix61_alt_mig.log -rw-r--r-- 1 root system 429109 May 18 14:33 lparaix53_alt_mig.log # tail –f lparaix53_alt_mig.log All rights reserved. US Government Users Restricted Rights - Use, duplication or disclosure restricted by GSA ADP Schedule Contract with IBM Corp. . . . . . << End of copyright notice for bos.help.msg.en_US >>. . . . Filesets processed: 566 of 2038 (Total time: 14 mins 7 secs). installp: APPLYING software for: bos.help.msg.en_US.com 220.127.116.11 # tail –f lparaix61_alt_mig.log All rights reserved. US Government Users Restricted Rights - Use, duplication or disclosure restricted by GSA ADP Schedule Contract with IBM Corp. . . . . . << End of copyright notice for bos.msg.Ja_JP >>. . . . Filesets processed: 418 of 2059 (Total time: 13 mins 17 secs). installp: APPLYING software for: bos.msg.JA_JP.rte 18.104.22.168
After the migration process is finished, the NIM master unmounts the cache file systems on the master and clients. We observe that the correct oslevel is returned at the end of the migration that is 7100-01.
At this point we restart each of the NIM clients into AIX 7.1. You'll observe that the bootlist for each client was modified by nimadm, so that the alternate rootvg is now the only disk in the boot list. With the clients successfully restarted on 7.1 the migration is now complete.
AIX Tunables post migration
After an AIX migration, I usually like to run the tuncheck command to verify the current tunable parameters are valid. One area that can indicate a tuning problem is the AIX error report. If you see the following messages in the errpt output, you might want to verify the current settings are valid:
IDENTIFIER TIMESTAMP T C RESOURCE_NAME DESCRIPTION D221BD55 0523115112 I O perftune RESTRICTED TUNABLES MODIFIED AT REBOOT --------------------------------------------------------------------------- LABEL: TUNE_RESTRICTED IDENTIFIER: D221BD55 Date/Time: Wed May 23 11:51:16 EET 2012 Sequence Number: 676 Machine Id: 00C342C64C00 Node Id: lparaix53 Class: O Type: INFO WPAR: Global Resource Name: perftune DescriptionRESTRICTED TUNABLES MODIFIED AT REBOOT Probable Causes SYSTEM TUNING User Causes TUNABLE PARAMETER OF TYPE RESTRICTED HAS BEEN MODIFIED Recommended Actions REVIEW TUNABLE LISTS IN DETAILED DATA Detail Data LIST OF TUNABLE COMMANDS CONTROLLING MODIFIED RESTRICTED TUNABLES AT REBOOT, SEE FILE /etc/tunables/lastboot.logvmo
In the output above, you'll notice that we are advised to check the /etc/tunables/lastboot.log for a modified restricted vmo tuning parameter. At this point I usually like to run the tuncheck command against the current /etc/tunables/nextboot file and review its output. As you can see, in the example below, we are warned that several restricted tunables are not set to their default values. These values might not be appropriate for your newly migrated AIX 7.1 (or 6.1) system. Settings that worked well with 5.3 are most likely no longer appropriate with 7.1.
Based on the output above, the tuning for this newly migrated 7.1 system appears to be inappropriate. Unless we have a valid reason (which has been verified by IBM AIX support) we should set these tunables to their default AIX 7.1 settings.
You can reset individual tunables to their defaults using the –d flag and the corresponding tuning command. For example to set the maxperm% tunable to its default you would run the following vmo command:
# vmo -p -d maxperm% Modification to restricted tunable maxperm%, confirmation required yes/no yes Setting maxperm% to 90 in nextboot file Setting maxperm% to 90 Warning: a restricted tunable has been modified
If you want to set all the vmo tunables back to their defaults you would run the following vmo command with the –D option:
# vmo -r -D Setting maxfree to 1088 in nextboot file Setting minfree to 960 in nextboot file Setting minperm% to 3 in nextboot file Modification to restricted tunable maxperm%, confirmation required yes/no yes Setting maxperm% to 90 in nextboot file Modification to restricted tunable strict_maxperm, confirmation required yes/no yes Setting strict_maxperm to 0 in nextboot file Setting maxpin% to 80 in nextboot file ...etc... Warning: some changes will take effect only after a bosboot and a rebootRun bosboot now? yes/no yes bosboot: Boot image is 47016 512 byte blocks.
Note that setting all parameters to their defaults will require the bosboot command to be run and a reboot of the system for the changes to take effect.
The tundefault command can also reset tuning to default parameters. The command launches all the tuning commands (ioo, vmo, schedo, no, nfso, and raso) with the -D flag. This resets all the AIX tunable parameters to their default values. The –r flag defers the reset to default value to the next reboot. This clears the stanza(s) in the /etc/tunables/nextboot file and if necessary, runs bosboot and warns that a reboot is needed.
# tundefault -r
After the tunables have been reset, re-run the tuncheck command and ensure it runs without errors:
# tuncheck -p -f /etc/tunables/nextboot Checking successful
You can verify when a systems tunables were lasted checked (with the tuncheck command) by reviewing the info stanza in the /etc/tunables/nextboot file. For example, the nextboot file (below) was last checked on 7th June 2012 and the system was running AIX 7.1.
Unless you've permanently set restricted tunables in your /etc/tunables/nexboot file, the migration will change the systems default tuning to match the newer version of AIX. For example, we observed the following tuning changes on our AIX 5.3 system after migrating to 7.1.
- The maxperm default value changed from 80 to 90:
maxperm% 80 80 80 1 100 % memory D maxperm% 90 90 90 1 100 % memory D
- The minperm default value changed from 20 to 3:
minperm% 20 20 20 1 100 % memory D minperm% 3 3 3 1 100 % memory D
- Note that with AIX 7.1 lru_file_repage is hardcoded to 0 and removed from the list of vmo tunables. Please refer to the following document, for more information.
Oracle Architecture and Tuning on AIX v2.20
- As expected, we noted a number of new tunables with AIX 7.1. Some examples are shown below.
vmm_klock_mode 2 2 2 0 3 numeric B j2_inodeCacheSize 200 200 200 1 1000 D
To learn more about these new tunables we can run the corresponding tuning utility with the –h flag to obtain usage information.
# vmo -h vmm_klock_mode
Help for tunable vmm_klock_mode:
Kernel locking prevents paging out kernel data. This improves system performance in many cases. If set to 0, kernel locking is disabled. If set to 1, kernel locking is enabled automatically if Active Memory Expansion (AME) feature is also enabled. In this mode, only a subset of kernel memory is locked. If set to 2, kernel locking is enabled regardless of AME and all of kernel data is eligible for locking. If set to 3, only the kernel stacks of processes are locked in memory. Enabling kernel locking has the most positive impact on performance of systems that do paging but not enough to page out kernel data or on systems that do not do paging activity at all. Note that 1, 2, and 3 are only advisory. If a system runs low on free memory and performs extensive paging activity, kernel locking is rendered ineffective by paging out kernel data. Kernel locking only has an impact on pageable page-sizes in the system.
Range: 0 - 3
If processes are being delayed waiting for compressed memory to become available, increase ame_minfree_mem to improve response time. Note, this must be at least 257kb less than ame_maxfree_mem.
- We also noticed a new subsystem, named aso. The Active System
Optimizer (ASO) is a new AIX 7.1 (on POWER7 only) feature that autonomously tunes the allocation of system resources to improve performance.
# lsitab -a | grep aso aso:23456789:once:/usr/bin/startsrc -s aso -e "NORMAL_RESPAWN_NOLOG"
- New network (no) tuning options also appeared after the migration. The example below relates to TCP tuning for loopback network access.
# no -a | grep tcp_fast tcp_fastlo = 0 tcp_fastlo_crosswpar = 0
# no -h tcp_fastlo
Help for tunable tcp_fastlo:
Specifies whether TCP fastpath loopback is enabled (1) or disabled (0).
Range: 0 - 1
This option allows the TCP loopback traffic to shortcut the entire TCP/IP stack (protocol and interface) in order to achieve better performances.
# no -h tcp_fastlo_crosswpar
Help for tunable tcp_fastlo_crosswpar:
Specifies whether TCP fastpath loopback between WPARs of a system is allowed (1) or forbidden (0).
Range: 0 - 1
This option is valid only if TCP fastpath loopback is enabled (with tcp_fastlo option).
I highly recommend that you refer to the IBM AIX Version 7.1 Differences Guide Redbook publication for more information on the new features of the AIX 7.1 OS.
Considerations when migrating to AIX 7.1 (or 6.1)
Do you have JFS file systems in rootvg? If you do, please be aware that nimadm does not convert rootvg file systems from JFS to JFS2. I have requested that IBM include this feature with nimadm in the future. Starting with AIX 6.1 TL4, the alt_disk_copy utility has a new –T flag to convert rootvg file systems to JFS2 during the cloning process. Unfortunately nimadm does not currently call this flag with alt_disk_copy.
If you are already running AIX 6.1 TL4 or higher, then you can use alt_disk_copy –T to convert rootvg file systems to JFS2 first and then use nimadm to migrate to AIX 7.1.
If you are on AIX 5.3, the alt_disk_copy command does not have the –T flag. In this case you might want to migrate to AIX 7.1 (or 6.1) with nimadm first, and then use the alt_disk_copy –T command to convert rootvg file systems to JFS2.
You might need to upgrade your MPIO device driver software. For example, if you are using IBM SDDPCM, then you will need to uninstall the previous version of SDDPCM and then install the correct version of the software for your new version of AIX. You might be able to use pre and post-migration scripts to include this update as part of the overall AIX migration process. Please refer to the following link for an example.
- As discussed in my previous nimadm article, multibos is not supported in nimadm environments. Before you start a nimadm migration make sure you have removed any old standby BOS instance and that your rootvg logical volumes are not using any bos_ LV names.
- During our tests we found that even though we removed the standy instance (multibos –R), the nimadm process failed with the following error:
+-----------------------------------------------------------------------------+ Executing nimadm phase 11. +-----------------------------------------------------------------------------+ Cloning altinst_rootvg on client, Phase 3. Client alt_disk_install command: alt_disk_copy -j -M 7.1 -P3 -d "hdisk1" ## Phase 3 ################### Verifying altinst_rootvg... alt_disk_copy: 0505-218 ATTENTION: init_multibos() returned an unexpected result. Cleaning up. forced unmount of /alt_inst/var/log forced unmount of /alt_inst/var/log forced unmount of /alt_inst/var forced unmount of /alt_inst/var forced unmount of /alt_inst/usr/local forced unmount of /alt_inst/usr/local forced unmount of /alt_inst/usr forced unmount of /alt_inst/usr forced unmount of /alt_inst/tmp forced unmount of /alt_inst/tmp forced unmount of /alt_inst/opt forced unmount of /alt_inst/opt forced unmount of /alt_inst/home forced unmount of /alt_inst/home forced unmount of /alt_inst/admin forced unmount of /alt_inst/admin forced unmount of /alt_inst forced unmount of /alt_inst 0505-187 nimadm: Error cloning altinst_rootvg on client. Cleaning up alt_disk_migration on the NIM master. Cleaning up alt_disk_migration on client lpar1. Client alt_disk_install command: alt_disk_install -M 7.1 -X Bootlist is set to the boot disk: hdisk0 blv=hd5
We also found the init_multibos error in the /var/adm/ras/alt_disk_inst.log file on the NIM client:
Tue Nov 22 14:48:12 EETDT 2011 cmd: /ALT_MIG_SPOT/sbin/alt_disk_copy -j -M 7.1 -P3 -d hdisk1 Verifying altinst_rootvg... alt_disk_copy: 0505-218 ATTENTION: init_multibos() returned an unexpected result. Cleaning up.
Given that the error appeared to be related to init_multibos, we assumed the failure was due to some multibos checks being performed by alt_disk_copy on the client. The client system did not have an existing multibos standby instance. So, we tried two things: First we created a standby instance on the client (multibos –s –X) and re-tried the nimadm operation. This failed. Next we removed the standby instance (multibos –R) and re-tried the nimadm operation. This worked and the client then migrated to AIX 7.1 successfully. We re-tried the same operations (that is, create standby instance, remove standby instance and nimadm) several times and each worked as expected.
Unfortunately, it appears that 'multibos –R' may not clean up the /bos_inst directory. If this directory exists the nimadm operation will most likely fail. The simple fix was (in our case) to remove the /bos_inst directory before attempting the AIX migration.
# rm –r /bos_inst
- If you plan on using your AIX 7.1 NIM master to migrate your AIX 5.3 clients to AIX
6.1, then make sure that you install the AIX 7.1 bos.alt_disk_install.rte fileset into the AIX 6.1 SPOT resource first. Failure to do so will result in your nimadm operation reporting the following error message:
You must install the AIX 7.1 bos.alt_disk_install.rte fileset into your AIX 6.1 SPOT resource.
# smit nim_res_op ....etc... > spotaix610605 Customize a SPOT Type or select values in entry fields. Press Enter AFTER making all desired changes. [Entry Fields] * Resource Name spotaix610605 * Source of Install Images [lpp_sourceaix710104] + Fileset Names [bos.alt_disk_install.rte]
You can verify the correct fileset is installed in your 6.1 SPOT using the following nim command:
# nim -o showres spotaix610605 | grep bos.alt_disk_install.rte bos.alt_disk_install.rte22.214.171.124 A F Alternate Disk Installation
Some administrators, migrating from AIX 5.3 to 6.1, reported issues with certain device filesets left behind after the migration. They identified there was an issue when the lppchk command returned errors after migration. To resolve the problem, IBM support advised that certain filesets be uninstalled and that several entries be deleted from the ODM. Please refer to the following link for further information.
Another 6.1 related issue we discovered was that after migrating, the sys0maxuproc attribute was returned to its default value. This resulted in performance issues and application crashes. Check your maxuproc value before and after the migration and ensure it has not changed. We did not experience this issue with AIX 7.1 migrations. Please refer to the following blog post for more information.
Read the latest AIX 7.1 installation tips document when planning your migrations. It can help you resolve and/or avoid issues either before, during or after the upgrade. For example:
When applying the 7100-01 Technology Level with Service Pack 1 included, you will have to run smitty update_all a second time to update bos.aso and mcr.rte. Until this is done, the oslevel command will not indicate the correct level.
- I recommend you migrate to the latest TL and SP for AIX 7.1 (or 6.1) to avoid any known issues. For example, SP3 for 7.1 TL1 is vulnerable to an issue where the netstat –f command can crash an LPAR. SP4 for 7.1 TL1 contains APAR IV09942 which resolves this problem. We hit this problem during our migrations. One of our system monitoring tools was regularly running the netstat command, which resulted in a number of unwanted system outages on some systems.
Customers using IBM DB2® Versions 8.2, 9.1, 9.5 or 9.7 should apply the following ifixes when applying SP4 for AIX 7.1 TL1 or AIX 6.1 TL7. Use APAR IV22132 for AIX 7.1 TL1 SP4 and APAR IV22062 for AIX 6.1 TL7 SP4. Please refer to the following website for further information:
Before applying AIX 7.1 TL1 SP4 to your system, make sure you have removed any legacy tuning from your system. During our tests we discovered that if we left legacy AIX 5.2 or 5.3 tuning in place the LPAR would hang during restart at "Setting tunable parameters". The following entries in our nextboot file were removed and the system started OK. This issue only appeared with systems migrated to SP4. Systems on AIX 7.1 TL1 SP3 did not exhibit this problem however we still removed the legacy tuning as it was inappropriate for 7.1.
info: AIX_level = "126.96.36.199" Kernel_type = "MP64" Last_validation = "2004-12-08 13:48:10 EETDT (current, reboot)" vmo: lrubucket = "262144" strict_maxclient = "1" strict_maxperm = "1" lru_file_repage = "0" minperm% = "2" maxperm% = "5" maxclient% = "5" maxfree = "1200" ioo: aio_maxreqs = "12288" j2_maxPageReadAhead = "512" maxpgahead = "8" no: use_isno = "0" udp_sendspace = "65536" udp_recvspace = "65536" tcp_sendspace = "262144" tcp_recvspace = "262144" rfc132
You might want to simply move the existing nextboot file "out of the way" prior to migrating. You can then review the file post migration. For example:
# mv /etc/tunables/nextboot /etc/tunables/nextboot.old
- For special considerations for SSH host keys and AIX migrations, refer to this blog.
Note: As of AIX Version 7.1, the 32-bit kernel has been deprecated. Therefore, 64-bit hardware is required to run AIX Version 7.1 (IBM POWER4™, POWER5™, POWER6™ or POWER7 systems only).
- Migrating to AIX 6.1 with nimadm
- IBM AIX Version 7.1 Differences Guide
- IBM AIX Version 6.1 Differences Guide
- AIX 7.1 Installation Tips (Doc Number=5565)
- 7100-01 Update When Using Cluster Aware AIX
- Upgrading to AIX 7
- Migrating to AIX 7.1 with nimadm via mksysb
- Convert rootvg file systems to JFS2 using alt_disk_copy
- nimadm - multibos error!?
- AIX 6.1 migration: iostat and maxuproc change to their defaults?
- An AIX Migration Tip Leads the Grab Bag
- tuncheck Command
- tundefault Command
- tunsave Command