Paging device in AIX

Hi All,

I want to know how to deactivate/delete  the paging device on the AIX system. I know some of the steps but just want to clarify the same. Is it goes like this :

# lsps -a
Page Space      Physical Volume   Volume Group    Size %Used Active  Auto  Type
paging00        hdisk474          rootvg       49152MB     1   yes   yes    lv
hd6             hdisk472          rootvg       49152MB     1   yes   yes    lv

#chps -a n /dev/paging00
#swapoff /dev/paging00
#rmlv /dev/paging00

Actually the thing is I want to use that disk hdisk474 for creating alternate disk installation, as I want to prepare this server for TL12 upgrade. Also, let me know what will be the steps to create the paging space again. kindly verify whether the above steps are correct ?

Who is Participating?
woolmilkporcConnect With a Mentor Commented:
Hi again,

your steps are basically correct, but don't use "rmlv"!
There is a special command "rmps" for that purpose!
It will take care to remove the PS definition from ODM.

Further, use the LV name, not the device name (i.e. omit "dev/...")!

So it's

chps -a n paging00   (not really needed, but can't do any harm)
swapoff paging00
rmps paging00

Creating a new pagespace is easy:

mkps -a -n -s number_of_partitions -t lv rootvg

Replace number_of_partitions with the desired value. Change rootvg to the name of a different VG if you don't want to use rootvg.

-a : activate at system restart
-n : also activate immediately
-s  number : number of PPs to use.
-t lv : (default) create a logical volume, not an NFS page space.

Please note that you can't give a name to your new paging space. It will be created automatically as paging00


Kerem ERSOYConnect With a Mentor PresidentCommented:

I understand that this isi the only paging space in your system. Paging space has two uses. The first is regular paging the econd is the dump area. So if you'll be changing the only paging space then you should chage the default dump dev to it too.

I don't know about disk layour but this volume might be in use for other volumes in rootvg too. Then you'll need to move all these volumes to other disks and remove the disk from the rootvg.

So the steps would be:

mkps -s 96 -a rootvg

issue sysdumpdev -l  and if it points to /dev/paging00 then issue this command:

sysdumpdev -P -p /dev/sysdumpnull

edit the /sbin/rc.boot and set the paging to the new name (the name would be something like /dev/paging01

change "swapon /dev/paging00" to "swapon /dev/paging01"

Deactivate the volume on next reboot
chps -a n /dev/pagigin00


lspv -l   hdisk474

make sure that no volume uses the disk


reducevg -d -f rootvg /dev/paging00

now you've removed the sdisk and moved the ps to another logical volume. Reenable sysdump:

sysdumpdev -P -p /dev/paging01


There is still hd6, isn't it?

Another tip: You can use "smitty mkps"  to create a new PS.

"smitty rmps" exists as well, but won't help you much, because you will have to run "smitty chps" and "smitty swapoff" beforehand. Using commandline is probably faster.

Free Tool: IP Lookup

Get more info about an IP address or domain name, such as organization, abuse contacts and geolocation.

One of a set of tools we are providing to everyone as a way of saying thank you for being a part of the community.

virgo0880Author Commented:
Ok got it, yes wmp is right, that hd6 is available.

One more thing is I have 2 disks - hdisk472 and hdisk473 in rootvg and I have 2 spare disks hdisk0 and hdisk1, all are of same sizes. Now, I dont want to break the mirror of rootvg and start the alternate disk installation, so whether I have to specify both hdisk0 and hdisk1 in the disk column of smitty or I should specify only one disk hdisk0. Correct me if I am wrong. If I specify hdisk0 & hdisk1, then whether alternate disk installation will take more time as it has to do the alternate install of the mirrored rootvg to 2 disks. Just want to clarify this point ?

No, you don't have to specify both disks.

But it won't take considerably longer anyway, because there is virtually no difference in creating an LV with one resp. two copies, and the actual writes will be carried out in parallel.

Please keep in mind that hdisk0 and hdisk1 must not belong to a VG!


virgo0880Author Commented:
You mean to say that as currently rootvg is having 2 on hdisk472 and mirrored on hdisk473. Now when I start alternate disk installation, i will specify hdisk0 and it will the create a cloned rootvg on hdisk0...right. That means, it will consider the mirrored rootvg(2 disks) as a single copy and create the clone on hdisk0...right ?

Yes.  A new VG will be created, LVs and filesystems will be defined and populated by means of "backup" and "restore", which is independent of whether the source LVs/FSs are mirrored or not.

Attention: Only mounted filesystems will be backed up, so take care not to have unmounted filesystems which you want to get cloned.
On the other hand, you can umount filesystems which you don't want to get cloned!

The above is true for cloning. Things are a bit different when restoring an mksysb. To get rid of mirroring here you will have to edit the file (COPIES=1) beforehand.

virgo0880Author Commented:
Just to clarify, I logged a call with IBM for this and they told that you have to specify 2 disks for cloning the mirrored rootvg (2 disks), also they commented saying that usually you should not do this procedure as it might create the problems and may be hung. So, always do alt disk install by unmirroring the root disk. What are your thoughts on this.

That's more a question for our other thread!

OK, if you got this info from IBM - what am I supposed to say?

I told you my thoughts in my previous comment -  but rather follow IBM, it might well be that they made experiences which I didn't.

virgo0880Author Commented:
Will update the status
virgo0880Author Commented:
Thanks for providing help on this.
Question has a verified solution.

Are you are experiencing a similar issue? Get a personalized answer when you ask a related question.

Have a better answer? Share it in a comment.

All Courses

From novice to tech pro — start learning today.