Re: Need a little help with AIX 4.3.3 and VG limitations




Thanks for the reply. I'll have a closer look at the migration steps you've mentioned below.
For reference the two primary hdisks I have for oracle datfiles are shown below.
Not sure if there is a preferred # of drives in an array or not
(ie. 2 - 10 disk RAID-10 array versus 5 - 4 disk RAID-10 arrays).
I'd like to add space to them by moving from the 18GB disks to the 36GB disks but I can't
add them all based on the number of drives I have now (28) and the fact I'm trying to add
another 16 which would take me over the 32 PV limit. That's why I was looking at the chvg -B option.
My other thought was I'd add say 4-36GB disks (RAID-10) and migrate some data there, remove a
few 18GB pdisks afterwards etc. create a new VG for exports, ... I'll figure something out.

RAID status.
hdisk2 03CF42BEDE7D4CO good 91.1GB RAID-10 array
hdisk3 03CF42BEDF154CO good 91.1GB RAID-10 array
List member disks of hdisk2.
pdisk0 0040AA25762000D member present 18.2GB Physical disk
pdisk5 0040AA25770100D member present 18.2GB Physical disk
pdisk1 0040AA25762100D member present 18.2GB Physical disk
pdisk6 0040AA2579A200D member present 18.2GB Physical disk
pdisk2 0040AA25769600D member present 18.2GB Physical disk
pdisk7 006094A717A600D member present 18.2GB Physical disk
pdisk3 0040AA25769A00D member present 18.2GB Physical disk
pdisk8 006094A717B400D member present 18.2GB Physical disk
pdisk4 0040AA2576F000D member present 18.2GB Physical disk
pdisk9 006094A7188500D member present 18.2GB Physical disk
List member disks of hdisk3.
pdisk16 000629D5377B00D member present 18.2GB Physical disk
pdisk22 0040AA22708500D member present 18.2GB Physical disk
pdisk17 000629D53D3600D member present 18.2GB Physical disk
pdisk23 0040AA22802300D member present 18.2GB Physical disk
pdisk18 000629D53FE000D member present 18.2GB Physical disk
pdisk24 0040AA22820100D member present 18.2GB Physical disk
pdisk19 000629D5409200D member present 18.2GB Physical disk
pdisk25 0040AA22820200D member present 18.2GB Physical disk
pdisk21 0040AA2191AC00D member present 18.2GB Physical disk
pdisk26 0040AA24373900D member present 18.2GB Physical disk

jbarratt@xxxxxxxxxxx 19/Jan/2007 6:14:39 pm >>>

1016*PP size=largest drive supported in the VG
1016*128=130048(larger than any drives mentioned) is the largest drive size you can add to the VG without changing tfactor (chvg -t)

if you change the tfactor, the Max PVs will be reduced. you dont need to do this.

you can obviously chvg -B(if you need to have more than 32 drives in the VG), but this may reduce the performance of the administrative commands, but should not impact the performance of the LVs in the VG.


you mention you have poor performance on 2 HDISKs that are mirrrored (RAID 10<>RAID 0+1) and (striping 0 has no impact on a two disk mirrored RAID set, so lets just call them RAID 1), or maybe I am misinterpreting what your issue is.

you can use migratepv or mklvcopy (which lies under migratepv) to increase the number of spindles that the LVs in question reside upon.

no downtime/backup-restore is necessary to migrate the LVs in question.

you need to determine how many spindles you want the LVs/VG to reside across.

the basic migration procedure is:

1. determine what disks the LV to be migrated lives on (lslv -m lvname), assuming it's mirrored properly, lets call them source1disks and source2disks. each column is a copy
2. determine what disks you intend to expand across, lets call them dest1disks and dest2disks
3. expand the existing vg so that the destination disks are included (extendvg myvg dest1disks dest2disks)
4. now make a third disk copy by LV with mklvcopy -k mylvname 3 dest1disks (you can drop the -k and syncvg -l mylvname if you prefer) this could take a very long time based on the number/size of PPs to migrate
5. when there are no longer any stale PPs, use rmlvcopy mylvname 2 source1disks
6. mklvcopy -k mylvname 3 dest2disks (you can drop the -k and syncvg -l mylvname if you prefer)
7. when there are no longer any stale PPs, use rmlvcopy mylvname 2 source2disks

that's the basics. others will have other methods, but this is the most basic way to move LVs












-----Original Message-----
From: IBM AIX Discussion List on behalf of Mark Schlechte
Sent: Fri 1/19/2007 5:26 PM
To: aix-l@xxxxxxxxxxxxx
Subject: Need a little help with AIX 4.3.3 and VG limitations


I have a 7026-6F1 server with 12x18GB internal SSA disks and an external D40 with 16x18GB disks.
This is an Oracle Database server that reports High IO rates some times on the 2 hdisks I have configured as RAID10 for datafiles.
I'm running out of space so I am planning on growing that storage by utilizing another D40 I have with 16-36GB disks.
I'm just trying to recall my limitations with respect to PP size and # of physical disks etc.


I'd like to just add this extra space to my existing 'non-rootvg' but if I check that VG I see:
VOLUME GROUP: myvg VG IDENTIFIER: 000203cfb9985aca
VG STATE: active PP SIZE: 128 megabyte(s)
VG PERMISSION: read/write TOTAL PPs: 2166 (277248 megabytes)
MAX LVs: 256 FREE PPs: 168 (21504 megabytes)
LVs: 32 USED PPs: 1998 (255744 megabytes)
OPEN LVs: 32 QUORUM: 5
TOTAL PVs: 8 VG DESCRIPTORS: 8
STALE PVs: 0 STALE PPs: 0
ACTIVE PVs: 8 AUTO ON: yes
MAX PPs per PV: 1016 MAX PVs: 32

So I see MAX PV's=32 and I'm already at 12+16=28.


If I recall there was some discussion previously about using the chvg with the -B option as in:
"-B Changes the volume group to big vg format. This can accommodate up to 128
physical volumes and 512 logical volumes."
I believe there were some mixed reviews here.


Otherwise depending on how my disk is configured (RAID types) I might be able to migrate some data
via migratepv if I could add some 36GB disks and make a few smaller RAID10 hdisks to stage to?
I'm trying to do this without any downtime or going through a big backup/restore process if possible.


Sorry for the general ramble but if something obvious jumps out at someone let me hear it.


TIA, Mark



<html><body><font face = "arial" size = "1">
DISCLAIMER: The information transmitted is intended only
for the addressee and may contain confidential,
proprietary and/or privileged material. Any
unauthorized review, distribution or other use
of or the taking of any action in reliance upon
this information is prohibited. If you received
this in error, please contact the sender and
delete or destroy this message and any copies.
</font><body></html>




<html><body><font face = "arial" size = "1">
DISCLAIMER: The information transmitted is intended only
for the addressee and may contain confidential,
proprietary and/or privileged material. Any
unauthorized review, distribution or other use
of or the taking of any action in reliance upon
this information is prohibited. If you received
this in error, please contact the sender and
delete or destroy this message and any copies.
</font><body></html>

Relevant Pages

  • Re: Need a little help with AIX 4.3.3 and VG limitations
    ... you can obviously chvg -B(if you need to have more than 32 drives in the VG), but this may reduce the performance of the administrative commands, but should not impact the performance of the LVs in the VG. ... you can use migratepv or mklvcopy to increase the number of spindles that the LVs in question reside upon. ... I'm running out of space so I am planning on growing that storage by utilizing another D40 I have with 16-36GB disks. ...
    (AIX-L)
  • Re: Problems replacing two bad SSA disks, bad vendor support, down system and trying to restore
    ... for defined drives that aren't being used you don't have to go into ... rmdev the old unused disk slots, and mkdev p & hdisks, mapping as ... > back the LV config for that vg by deleting the two failed disks. ... force a break of a mirror, remove the dead drives, add the new ones (even ...
    (comp.unix.aix)
  • Some Test with DVD-RAM
    ... I have done some quality assessment of two DVD-RAM capable ... DVD-burners with four different DVD-RAM brands. ... both drives ... These disks are incredibly bad. ...
    (comp.sys.ibm.pc.hardware.storage)
  • RE: bad blocks... random death
    ... > the disks with bad blocks work actually fine elswhere (in ... The drives are fine. ... including your drive cables, ... grounded to the chassis and through the chassis to the ...
    (RedHat)
  • Re: Need Recovery Disk?
    ... It is, and this is, of course, the fatal flaw in the recovery partition ... to immediately create restore disks.. ... reproduced the drives in the first place. ... support for a laptop owned by a client, a US citizen who purchased the unit ...
    (microsoft.public.windowsxp.help_and_support)