Topics: AIX, Backup & restore, System Admin
How to restore an image.data file from an existing mksysb file
Change the /tmp directory (or a directory where you would like to store the /image.data file from the mksysb image) and restore the /image.data file from the mksysb:
If you want to list the files in a mksysb image first, you can run the following command:# cd /tmp # restore -xqvf [/location/of/mksysb/file] ./image.data
# restore -Tqvf [/location/of/mksysb/file]
Create a new image.data file by running the following command:
Edit the image.data file to break the mirror, by running the following command:# cd / # mkszfile
# vi /image.dataWhat you are looking for are the "lv_data" stanzas. There will be one for every logical volume associated with rootvg.
The following is an example of an lv_data stanza from an image.data file of a mirrored rootvg. The lines that need changing are marked bold:
lv_data:Note: There are two disks in the 'LV_SOURCE_DISK_LIST', THE 'COPIES' value reflects two copies, and the 'PP' value is double that of the 'LPs' value.
VOLUME_GROUP= rootvg
LV_SOURCE_DISK_LIST= hdisk0 hdisk1
LV_IDENTIFIER= 00cead4a00004c0000000117b1e92c90.2
LOGICAL_VOLUME= hd6
VG_STAT= active/complete
TYPE= paging
MAX_LPS= 512
COPIES= 2
LPs= 124
STALE_PPs= 0
INTER_POLICY= minimum
INTRA_POLICY= middle
MOUNT_POINT=
MIRROR_WRITE_CONSISTENCY= off
LV_SEPARATE_PV= yes
PERMISSION= read/write
LV_STATE= opened/syncd
WRITE_VERIFY= off
PP_SIZE= 128
SCHED_POLICY= parallel
PP= 248
BB_POLICY= non-relocatable
RELOCATABLE= yes
UPPER_BOUND= 32
LABEL=
MAPFILE= /tmp/vgdata/rootvg/hd6.map
LV_MIN_LPS= 124
STRIPE_WIDTH=
STRIPE_SIZE=
SERIALIZE_IO= no
FS_TAG=
DEV_SUBTYP=
The following is an example of the same lv_data stanza after manually breaking the mirror. The lines that have been changed are marked bold. Edit each 'lv_data' stanza in the image.data file as shown below to break the mirrors.
lv_data:Note: The 'LV_SOURCE_DISK_LIST' has been reduced to one disk, the 'COPIES' value has been changed to reflect one copy, and the 'PP' value has been changed so that it is equal to the 'LPs' value.
VOLUME_GROUP= rootvg
LV_SOURCE_DISK_LIST= hdisk0
LV_IDENTIFIER= 00cead4a00004c0000000117b1e92c90.2
LOGICAL_VOLUME= hd6
VG_STAT= active/complete
TYPE= paging
MAX_LPS= 512
COPIES= 1
LPs= 124
STALE_PPs= 0
INTER_POLICY= minimum
INTRA_POLICY= middle
MOUNT_POINT=
MIRROR_WRITE_CONSISTENCY= off
LV_SEPARATE_PV= yes
PERMISSION= read/write
LV_STATE= opened/syncd
WRITE_VERIFY= off
PP_SIZE= 128
SCHED_POLICY= parallel
PP= 124
BB_POLICY= non-relocatable
RELOCATABLE= yes
UPPER_BOUND= 32
LABEL=
MAPFILE= /tmp/vgdata/rootvg/hd6.map
LV_MIN_LPS= 124
STRIPE_WIDTH=
STRIPE_SIZE=
SERIALIZE_IO= no
FS_TAG=
DEV_SUBTYP=
Save the edited image.data file. At this point you can use the edited image.data file to do one of the following: You can now use your newly edited image.data file to create a new mksysb to file, tape, or DVD.
E.g.: To file or tape: place the edited image.data file in the / (root) directory and rerun your mksysb command without using the "-i" flag. If running the backup through SMIT, make sure you set the option "Generate new /image.data file?" to 'no' (By default it is set to 'yes').
To DVD: Use the -i flag and specify the [/location] of the edited image.data file. If running through SMIT specify the edited image.data file location in the "User supplied image.data file" field.
Within NIM you would create an 'image_data' resource for use with NIM to restore a mksysb without preserving mirrors.
Note: If you don't want to edit the image.data file manually, here's a script that you can use to have it updated to a single disk for you, assuming your image_data file is called /image.data:
cat /image.data | while read LINE ; do
if [ "${LINE}" = "COPIES= 2" ] ; then
COPIESFLAG=1
echo "COPIES= 1"
else
if [ ${COPIESFLAG} -eq 1 ] ; then
PP=`echo ${LINE} | awk '{print $1}'`
if [ "${PP}" = "PP=" ] ; then
PPNUM=`echo ${LINE} | awk '{print $2}'`
((PPNUMNEW=$PPNUM/2))
echo "PP= ${PPNUMNEW}"
COPIESFLAG=0
else
echo "${LINE}"
fi
else
echo "${LINE}"
fi
fi
done > /image.data.1disk
Topics: AIX, Backup & restore, NIM, System Admin↑
Creating an image_data resource without preserving mirrors for use with NIM
Transfer the /image.data file to the NIM master and store it in the location you desire. It is a good idea to place the file, or any NIM resource for that matter, in a descriptive manor, for example: /export/nim/image_data. This will ensure you can easily identify your "image_data" NIM resource file locations, should you have the need for multiple "image_data" resources.
Make sure your image.data filenames are descriptive also. A common way to name the file would be in relation to your clientname, for example: server1_image_data.
Run the nim command, or use smitty and the fast path 'nim_mkres' to define the file that you have edited using the steps above:
From command line on the NIM master:
# nim -o define -t image_data -a server=master -a location=/export/nim/image_data/server1_image_data -a comments="image.data file with broken mirror for server1" server1_image_dataNOTE: "server1_image_data" is the name given to the 'image_data' resource.
Using smit on the NIM master:
# smit nim_mkresSelect 'image_data' as the Resource Type. Then complete the following screen:
Define a Resource
Type or select values in entry fields.
Press Enter AFTER making all desired changes.
[Entry Fields]
* Resource Name [server1_image_data]
* Resource Type image_data
* Server of Resource [master]
* Location of Resource [/export/nim/image_data/server1_image_data]
Comments []
Source for Replication []
Run the following command to make sure the 'image_data' resource was created:
# lsnim -t image_dataThe command will give output similar to the following:
Run the following command to get information about the 'image_data' resource:# lsnim -t image_data server1_image_data resources image_data
# lsnim -l server1_image_data server1_image_data: class = resources type = image_data Rstate = ready for use prev_state = unavailable for use location = /export/nim/image_data/server1_image_data alloc_count = 0 server = master
Topics: AIX, Backup & restore, NIM, System Admin↑
Using the image_data resource to restore a mksysb without preserving mirrors using NIM
Specify using the 'image_data' resource when running the 'bosinst' command from the NIM master:
From command line on the NIM master:
# nim -o bos_inst -a source=mksysb -a lpp_source=[lpp_source] -a spot=[SPOT] -a mksysb=[mksysb] -a image_data=mksysb_image_data -a accept_licenses=yes server1Using smit on the NIM master:
# smit nim_bosinstSelect the client to install. Select 'mksysb' as the type of install. Select a SPOT at the same level as the mksysb you are installing. Select an lpp_source at the same level than the mksysb you are installing.
NOTE: It is recommended to use an lpp_source at the same AIX Technology Level, but if using an lpp_source at a higher level than the mksysb, the system will be updated to the level of the lpp_source during installation. This will only update Technology Levels.
If you're using an AIX 5300-08 mksysb, you cannot use an AIX 6.1 lpp_source. This will not migrate the version of AIX you are running to a higher version. If you're using an AIX 5300-08 mksysb and allocate a 5300-09 lpp_source, this will update your target system to 5300-09.
Install the Base Operating System on Standalone Clients
Type or select values in entry fields.
Press Enter AFTER making all desired changes.
[TOP] [Entry Fields]
* Installation Target server1
* Installation TYPE mksysb
* SPOT SPOTaix53tl09sp3
LPP_SOURCE [LPPaix53tl09sp3]
MKSYSB server1_mksysb
BOSINST_DATA to use during installation []
IMAGE_DATA to use during installation [server1_image_date]Topics: AIX, Backup & restore, NIM, System Admin↑
How to unconfigure items after mksysb recovery using NIM
There will be a situation where you want to test a mksysb recovery to a different host. The major issue with this is, that you bring up a server within the same network, that is a copy of an actual server that's already in your network. To avoid running into 2 exactly the same servers in your network, here's how you do this:
First make sure that you have a separate IP address available for the server to be recovered, for configuration on your test server. You definitely don't want to bring up a second server in your network with the same IP configuration.
Make sure you have a mksysb created of the server that you wish to recover onto another server. Then, create a simple script that disables all the items that you don't want to have running after the mksysb recovery, for example:
# cat /export/nim/cust_scripts/custom.ksh
#!/bin/ksh
# Save a copy of /etc/inittab
cp /etc/inittab /etc/inittab.org
# Remove unwanted entries from the inittab
rmitab hacmp 2>/dev/null
rmitab tsmsched 2>/dev/null
rmitab tsm 2>/dev/null
rmitab clinit 2>/dev/null
rmitab pst_clinit 2>/dev/null
rmitab qdaemon 2>/dev/null
rmitab sddsrv 2>/dev/null
rmitab nimclient 2>/dev/null
rmitab nimsh 2>/dev/null
rmitab naviagent 2>/dev/null
# Get rid of the crontabs
mkdir -p /var/spool/cron/crontabs.org
mv /var/spool/cron/crontabs/* /var/spool/cron/crontabs.org/
# Disable start scripts
chmod 000 /etc/rc.d/rc2.d/S01app
# copy inetd.conf
cp /etc/inetd.conf /etc/inetd.conf.org
# take out unwanted items
cat /etc/inetd.conf.org | grep -v bgssd > /etc/inetd.conf
# remove the hacmp cluster configuration
if [ -x /usr/es/sbin/cluster/utilities/clrmclstr ] ; then
/usr/es/sbin/cluster/utilities/clrmclstr
fi
# clear the error report
errclear 0
# clean out mail queue
rm /var/spool/mqueue/*
# smitty nim_mkresSelect 'script' and complete the form afterwards. For example, if you called it 'UnConfig_Script':
Then, when you are ready to perform the actual mksysb recovery using "smitty nim_bosinst", you can add this script resource on the following line:# lsnim -l UnConfig_Script UnConfig_Script: class = resources type = script comments = Rstate = ready for use prev_state = unavailable for use location = /export/nim/cust_scripts/custom.ksh alloc_count = 0 server = master
Customization SCRIPT to run after installation [UnConfig_Script]
Topics: AIX, EMC, SAN, Storage, System Admin↑
Unable to remove hdiskpower devices due to a method error
If you get a method error when trying to rmdev -dl your hdiskpower devices, then follow this procedure.
Cannot remove hdiskpower devices with rmdev, get error "method error (/etc/methods/ucfgpowerdisk):"The fix is to uninstall/reinstall Powerpath, but you won't be able to until you remove the hdiskpower devices with this procedure:
# odmdelete -q name=hdiskpowerX -o CuDv
(for every hdiskpower device)# odmdelete -q name=hdiskpowerX -o CuAt
(for every hdiskpower device)# odmdelete -q name=powerpath0 -o CuDv
# odmdelete -q name=powerpath0 -o CuAt
# rm /dev/powerpath0
- You must remove the modified files installed by powerpath and then reboot the server. You will then be able to uninstall powerpath after the reboot via the "installp -u EMCpower" command. The files to be removed are as follows:
(Do not be concerned if some of the removals do not work as PowerPath may not be fully configured properly).# rm ./etc/PowerPathExtensions # rm ./etc/emcp_registration # rm ./usr/lib/boot/protoext/disk.proto.ext.scsi.pseudo.power # rm ./usr/lib/drivers/pnext # rm ./usr/lib/drivers/powerdd # rm ./usr/lib/drivers/powerdiskdd # rm ./usr/lib/libpn.a # rm ./usr/lib/methods/cfgpower # rm ./usr/lib/methods/cfgpowerdisk # rm ./usr/lib/methods/chgpowerdisk # rm ./usr/lib/methods/power.cat # rm ./usr/lib/methods/ucfgpower # rm ./usr/lib/methods/ucfgpowerdisk # rm ./usr/lib/nls/msg/en_US/power.cat # rm ./usr/sbin/powercf # rm ./usr/sbin/powerprotect # rm ./usr/sbin/pprootdev # rm ./usr/lib/drivers/cgext # rm ./usr/lib/drivers/mpcext # rm ./usr/lib/libcg.so # rm ./usr/lib/libcong.so # rm ./usr/lib/libemcp_mp_rtl.so # rm ./usr/lib/drivers/mpext # rm ./usr/lib/libmp.a # rm ./usr/sbin/emcpreg # rm ./usr/sbin/powermt # rm ./usr/share/man/man1/emcpreg.1 # rm ./usr/share/man/man1/powermt.1 # rm ./usr/share/man/man1/powerprotect.1
- Re-install Powerpath.
How best to configure the /etc/netsvc.conf file, making it easier to troubleshoot when resolving DNS issues: This file should resolve locally and through DNS. The line would read:
hosts=local,bindYou then would need to make sure that all the local adapter IP addresses are entered in /etc/hosts. After that is complete, for every adapter on the system you would apply:
# hostThis will ensure a host command generates the same ouput (the hostname) with and without /etc/netsvc.conf. That way, you'll know you can continue to do certain things while troubleshooting a DNS problem.
Here are a couple of rules that your paging spaces should adhere to, for best performance:
- The size of paging space should match the size of the memory.
- Use more than one paging space, on different disks to each other.
- All paging spaces should have the same size.
- All paging spaces should be mirrored.
- Paging spaces should not be put on "hot" disks.
The command svmon -G can be used to determine the actual memory consumption of a server. To determine if the memory is over-committed, you need to divide the memory-virtual value by the memory-size value, e.g.:
# svmon -G
size inuse free pin virtual
memory 5079040 5076409 2631 706856 2983249
pg space 7864320 12885
work pers clnt other
pin 540803 0 2758 163295
in use 2983249 0 2093160
PageSize PoolSize inuse pgsp pin virtual
s 4 KB - 4918761 12885 621096 2825601
m 64 KB - 9853 0 5360 9853
In this example, the memory-virtual value is 2983249, and the memory-size value is 5079040. Note that the actual memory-inuse (5076409) is nearly the same as the memory-size (5079040) value. This is simply AIX caching as much as possible in its memory. Hence, the memory-free value is typically very low, 2631 in the example above. As such, determining the memory size based on the memory-free value does not provide a good interpretation of the actual memory consumption, as memory typically includes a lot of cached data.
Now, to determine the actual memory consumption, divide memory-virtual by memory-size:
Thus, the actual memory consumption is 58% of the memory. The size of the memory is 5079040 blocks of 4 KB = 19840 MB. The free memory is thus: (100% - 58%) * 19840 MB = 8332 MB.# bc scale=2 2983249/5079040 .58
Try to keep the value of memory consumption less than 90%. Above that, you will generally start seeing paging activity using the vmstat command. By that time, it is a good idea to lower the load on the system or to get more memory in your system.
No more ordering CDROMs or DVDs and waiting days. Download the .iso image over the web and install from there. Use the virtual DVD drive on your VIOS 2.1 and install directly into the LPAR or read the contents into your NIM Server.
Mount the .ISO image:
- On AIX 6.1 or AIX 7, use the loopmount command: http://publib.boulder.ibm.com/infocenter/aix/v6r1/index.jsp?topic=/com.ibm.aix.cmds/doc/aixcmds3/loopmount.htm
- On AIX 5.3, use the mklv-dd-mount trick: https://www.ibm.com/developerworks/wikis/display/WikiPtype/AIXV53MntISO
You have to prove you are entitled via: Customer number, Machine serial numbers or SWMA. The Entitled Software Support Download User Guide can be downloaded here: ftp://public.dhe.ibm.com/systems/support/planning/essreg/I1128814.pdf. Then you can download the AIX media, Expansion Packs, Linux Toolbox and more.
Start at: www.ibm.com/eserver/ess.
