Information about RAID LEVELS in HPUX

[P.PARTHIBAN]

hello all

I need a small clarification about Raid level 3 and Raid level 5
in hpux. If any disks fails in both Raid levels can we access the data during online without hotspare. for example one volume group called vg02 has been created in harddisks those are in
Raid level 3 or Raid level 5 . IF any harddisk fails in above
Raid level, will the volumegroup vg02 and its logical volumes
be working contineouly without any problem ?
Pls mention the differences between both Raid levels (3 and 5)
clearly. (performancewise)

Note: I am not using any hotspare in above mentioned example


Thanks and Regards

P.Parthiban

[Patrick Wessel]

RAID 5 is a block striped mode with data redundancy. Data is striped, using only the necessary disk mechs within the array to complete the transfer, but it will also allocate one additional mech in order to generate parity. Parity is not built onto a single disk, but will be located on different disks to allow I/O concuuency.

RAID 3 is a byte striped mode with data redundancy. Data is striped across the disks in a byte basis, and a single mech is used for parity. All mechs data and parity are accessed in parallel, and data is written in exactly the same location on all of the disks, keeping seek time and latency approximately identical throughout


RAID 3 is best for applications which have a large block transfers or sequential data transfers. RAID 5 is best for applications which have small random transfers, and are not more than 50% writes

[P.PARTHIBAN]

Dear PATRICK WESSEL

I couldn't find any answer from your reply for my first question?
i e will the volumegroup and its logical volumes work
contineously if any disk fails which is in Raid 3 or Raid 5
Pls clarify.

Thanks and Regards

P.Parthiban

[Evan Day]

The answer is yes, your volume group will continue to run normally. However, you will want to replace the faulty drive ASAP as you do run the risk of data loss should another drive fail.

[Venu_2]

Dear parthiban,

>>> If any disks fails in both Raid levels >>> can we access the data during online
>>> without hotspare.

Yes, u can access the data even if one disk fails.

This is how HOT SPARE works

When any disk fails in a healthy RAID 3 or 5 config. The RAID controller starts rebuilding the data on failed disk on to the HOT SPARE.

Once the defective disk is replaced data will be restored back to that and the HOT SPARE will be released.

Both raid level operates with data parity principal.

The only deference between is 3 and 5 is parity is stored in dedicated disk in 3 and is striped in 5.

Performance of the RAID 3 is mainly dependent on the amount of memory on the RAID controller. Go through the RAID manual for the detailed info.

regards

venu


. IF any harddisk fails in above
Raid level, will the volumegroup vg02 and its logical volumes
be working contineouly without any problem ?
Pls mention the differences between both Raid levels (3 and 5)
clearly. (performancewise)

Note: I am not using any hotspare in above mentioned example


Thanks and Regards

P.Parthiban

[Patrick Wessel]

Looks like I missed to explain an important point.
As I mentioned above do both (Raid 5 & Raid 3) store the parity information along with the data. If one drive fails is the RAID controller able to restore the data using this parity information. The outside world behind the RAID controller does not notices that one drive failed.
In RAID 3 and RAID 5 does the volumegroup not notice that the first drive failed, but after the second failed drive is the data lost.
As you mentioned has RAID and spare disk nothing to do with each other. But, if you like to, you cane combine it.

[Patrick Wessel]

Maybe you want to read the following white paper:
http://docs.hp.com/hpux/content/diskwp2.pdf

It's not up to date (August 1996), but basics never change