New to Integrity virtual machine - some queries

[Amit Jindal_1]

Hi,
I am new to IVM, just going through the documentstions and the forums questions. My existing project wants to switch from SD 32000 to Blade BL870C with IVM on top. I have a few queries:

1)The web sites / introduction pdf says sub cpu , shared i/o etc , but what are the limitations ?
2)If I have two HBA cards, can i allocate one to one guest O.S and second to other? If a HBA card is attached , where the LUNS will be visible ?
3) The requirement is to put around 2.5 TB spread across ( 500 LUNS )for a guest O.S ? so is there any limitations of LUN numbers allocated to any IVM guest host.

4)Your experience - Is the install /manegement of IVM complex, should i use the earlier Vpar which I am comfortable with ? i don't want to use sub cpu etc .. even if i can carve out 3 partitions on a BL870C i am fine. With earlier vpar i can allocate HBA cards to individual partitions, so management will be easy ..

[Torsten.]

All the limits are written in the documentation, for example

Maximum size of backing store (disk, lvol, file) 2 TB

virtual SCSI devices / VMâ Maximum 30

3.1.10 Configuration Limits
Table 3-2 lists the configuration limits for Integrity VM Version 3.5.

http://docs.hp.com/en/T2767-90105/ch03s01.html#configlimits


You spread 2.5TB over 500 LUNS? really? Sounds like overkill ...

BTW, vPars is not supported on blades ...

[Amit Jindal_1]

Thanks Torsten for the quick reply, I have checked the link. for one of the point it says ( virtual SCSI devices / VM â Maximum = 30 ) and ( Maximum size of backing store (disk, lvol, file)= 2 TB )

so what's the meaning of the first point ( sorry to ask a basic question : what's a SCSI device, a HBA controller port is a SCSI device or the LUNS attached to it are also classified as SCSI device )

if I am using a dual port HBA mezannine cards and attaching it to the array. The individual LUN size will be around 14GB ( Raw device for a DB ) and total around 2.5 TB = 2560 GB so around 183 LUNS , forcasted growth to around 300 LUNS, i gave a very high upper limit of 500.

So can't i hook it to HBA card with more than 30 LUNS or what .. ?

[Amit Jindal_1]

I have some info for help:

Existing setup : one SD32K npar 4CPU ( 2 vpars ) @ 2 cpu each ( 750 MHz PA8700 RISC , 2.5 TB atatched to one vpar and 1 TB attached to other vpar )

Planned : planned to be moved to BL870C blade with 2 guest O.Ses @ 2 cpu each. My worry is will is how i will allocate the LUNS ( planning to use 2 mezannine HBA cards )

The LUN size can't be changed ( project requirement ) 14 GB each LUN.

Does this crosses any IVM limits ?

[Tim Nelson · Accepted solution]

>>so what's the meaning of the first point ( sorry to ask a basic question : what's a SCSI device, a HBA controller port is a SCSI device or the LUNS attached to it are also classified as SCSI device )


You can only present 30 backingstores to each Guest.

If you plan to present 300 14GB devices separately to a guest it will not work. The number is 30 max.

Now you can present a large lun < 2TB to a Guest and then chop it up on the guest side with LMV.

[Amit Jindal_1]

Thanks Tim for the clarification, it will be problem then in my situation, these LUNS ( 14 GB X 300 around ) are refreshed LUNS from production , i.e clone of production DB in RAW format ( which is cloned on a staging server ) and presented to this DEV server for ST / DEV purposes.

I can't change the size as it will corrupt the DB data. So if i use a individual blade then I hope these limitations won't be there ?

[Amit Jindal_1]

thanks for the answers.