|
|||||||
| Forums | Search Forums | Register | Forum Rules | Man Pages | Albums | FAQ | Members | Calendar | Search | Today's Posts | Mark Forums Read |
| AIX AIX is IBM's industry-leading UNIX operating system that meets the demands of applications that businesses rely upon in today's marketplace. |
|
|
|
Thread Tools | Search this Thread | Display Modes |
|
#1
|
||||
|
||||
|
Position of the logical volume on the physical volume
Hello everyone,
I just read that while creating a logical volume(LV) we can choose the region of the physical volume (PV) in which the LV should be created. When I say region I mean: outer edge - outer middle - center - inner middle and inner edge. Can anyone help me understand the utility of that possibility ? How can we do that ( example please) ? Thank you |
| Sponsored Links | ||
|
|
#2
|
|||
|
|||
|
In smitty when you are creating a LV, you'll see: Code:
POSITION on physical volume middle You can change this to whatever you want. If you are using a SAN though, it probably won't make any difference since most modern SANs move data around as needed and spread it across multiple disks all unknown to the OS. If you are using internal physical disks, you will most likely see a difference - outer is faster than the inner because the head that reads the disk can read more on the outer edge because more disk is passed under it in a single disk rotation. |
| Sponsored Links | ||
|
|
#3
|
||||
|
||||
|
the center is the fastest on a disk - obviously this works only when you don't use raid 5 or SAN - you place it while you are creating the lv - either via smitty or with -a c option
Regards zxmaus |
|
#4
|
|||
|
|||
|
The center is fastest? Outer tracks would have more sectors so as the disk spins, the head is able to read more sectors in a single disk rotation, right?
---------- Post updated at 10:13 AM ---------- Previous update was at 10:06 AM ---------- Here is a descent animation of the head seeking data on Western Digital's website. It is advertising their "IntelliSeek", but it shows how the head reads data. Western Digital |
| Sponsored Links | |
|
|
#5
|
||||
|
||||
|
the outer edge of the disk has a faster rotational velocity than the slower inner edge and will whip around much more quickly on the outside than in the middle. However, because not all the data on the disk will be written to the outer edge as a result of the limitations of the physical placement of the data, the fastest seek times for the hard disk heads will be in the center area of the disk, where the head is most likely to pass on average.
|
| Sponsored Links | |
|
|
#6
|
|||
|
|||
|
I guess that is why UNIX gives the option of where you want the data because if you are worried about seek times, you'd want the data in the middle, if you were wanting raw speed to save a smaller number of large files, you'd be better off putting it on the outer edges... to each his own. Long live UNIX.
|
| Sponsored Links | |
|
|
#7
|
|||
|
|||
|
IMHO the problem is placement within regions based on application activity. The basic concept you want to follow: if the access is sequential in application terms the best I/O performance is achieved when the physical I/O access is sequential. When application I/O is random there are many different things to consider - one of them being - is the I/O still physical, or is it VMM cached, or application cached. In short, with modern systems (assuming more than one or two disks, i.e., separate volume groups for each application's data) there is no one best place. Assuming some physical relationship the only advice that remains is that AIX LVM mirrored logical volumes with lots of write activity may benefit from being placed on the outer edge because the VGSA - Volume Group Status Area - is on the outer edge and gets updated anytime a mirrored LP is modified. command you should look at: lslv - options: none, and with -l lspv -p Code:
michael@x054:[/home/michael]lslv lvDAV LOGICAL VOLUME: lvDAV VOLUME GROUP: vgData LV IDENTIFIER: 005d858f00004c0000000115722f3acc.15 PERMISSION: read/write VG STATE: active/complete LV STATE: closed/syncd TYPE: jfs2 WRITE VERIFY: off MAX LPs: 1024 PP SIZE: 16 megabyte(s) COPIES: 1 SCHED POLICY: parallel LPs: 1024 PPs: 1024 STALE PPs: 0 BB POLICY: relocatable INTER-POLICY: minimum RELOCATABLE: yes INTRA-POLICY: middle UPPER BOUND: 6 MOUNT POINT: N/A LABEL: None MIRROR WRITE CONSISTENCY: on/ACTIVE EACH LP COPY ON A SEPARATE PV ?: yes Serialize IO ?: NO INFINITE RETRY: no Code:
michael@x054:[/home/michael]lslv -l lvDAV lvDAV:N/A PV COPIES IN BAND DISTRIBUTION hdisk1 1024:000:000 0% 000:000:541:483:000 This looks "bad" or less than ideal - 0% and in two regions - however, in physical terms it is a sequential as it gets. A simple command will improve the 0% to >50% (see below). Code:
michael@x054:[/home/michael]lspv -p hdisk1 | grep lvDAV 2085-2625 used center lvDAV jfs2 N/A 2626-3108 used inner middle lvDAV jfs2 N/A Here, I can see the logical volume is organized sequentially - starting in the center region, and proceeding into the "inner middle" region as one long sequence. Now to get better stats I can do the following: (change the preferred region to center from middle) Code:
michael@x054:[/home/michael]chlv -a center lvDAV
0516-652 chlv: The -a parameter for IntraPolicy must be m, im, c,
ie or e.
michael@x054:[/home/michael]chlv -a c lvDAV
michael@x054:[/home/michael]lslv -l lvDAV
lvDAV:N/A
PV COPIES IN BAND DISTRIBUTION
hdisk1 1024:000:000 52% 000:000:541:483:000Hope this helps! |
| Sponsored Links | ||
|
![]() |
| Thread Tools | Search this Thread |
| Display Modes | |
More UNIX and Linux Forum Topics You Might Find Helpful
|
||||
| Thread | Thread Starter | Forum | Replies | Last Post |
| Confusion Regarding Physical Volume,Volume Group,Logical Volume,Physical partition | kashifsd17 | UNIX for Dummies Questions & Answers | 6 | 07-06-2012 04:19 PM |
| Logical volume name conflict in two volume group | dantares | AIX | 10 | 12-07-2009 12:12 PM |
| Unmount and remove all Logical vol.Volume group and physical disk | joeli | HP-UX | 3 | 01-27-2009 09:59 AM |
| Basic Filesystem / Physical Volume / Logical Volume Check | chipahoys | AIX | 1 | 12-02-2008 07:05 AM |
| LVM - Extending Logical Volume within Volume Group | ghimanshu | UNIX for Advanced & Expert Users | 3 | 07-26-2007 07:39 AM |
|
|