DS2300 (2 BCC), Host (2 HBA), Need sw?

[Jorge Pons]

Hi

Before: HPUX Box L1000 (1 HBA)- DS2300 (1BCC) with 14 disks. FULL BUS.
After: HPUX Box L1000 (2 HBA)- DS2300 (2BCC) with the same disks.

Do I need any Software to have High Availability?

Thanks in Advance

[Cheryl Griffin]

Search the user guide for High Availability

HP StorageWorks Disk System 2300 User's Guide
(PDF, 2.33MB)

http://h20000.www2.hp.com/bizsupport/TechSupport/DocumentIndex.jsp?contentType=SupportManual&locale=en_US&prodTypeId=12169&prodSeriesId=81148&taskId=115&prodSeriesId=81148&docIndexId=763

[Mohanasundaram_1]

Hi Jorge,

Have you put the Ds2300 in split bus mode? If not they will remain in the same bus having 2 controllers. Which obviously is not supported.

2 HBAs should have split bus mode. Then you can purchase the optional software product MirrorDisk/UX and mirror disks across the HBAs.

Pages 27-32 of the Ds2300 user's manual has some topology information.

Hope this helps.

With regards,
Mohan.

[D Block 2]

I'm assuming the ds2300 does not run in SPLIT bus mode (ok, maybe I am wrong, please confirm).

In an HA environment, having two links to an ds2300 is good. you eliminate the SinglePointOfFailure from the HOst L1000 server. If an scsi link card goes bad, you have an alternate..link.


I'll assume you are using LVM using PV-Links of some sort.

I'll assume you are using LVM using MIRROR'ing software.. if the boot disk goes bad, then you have an alternate disk to boot from..

we Logical Volume also have a duplicate (Mirror'ed logical volume as part of the DS2300).

here's, the point, do you want another DS2300 to provide yet another redundant h/w Jpod ?

how far do you want to take this ?

I'm not sure if you can split the ds2300 like the older scsi devies ..

best of luck.. would suggest you create yet another back-door using IGNITE to Tape.

[Jorge Pons]

Hi

The L1000 have internal disks with MirrorUX.
The DS2300 is in FULL BUS.

with "ioscan" I can see all the disks in both SCSI controllers, but in SAM, only see the disks in the first SCSI controller where it was configured. If I see "/etc/lvmtab" I only see the "vgxx" with one disk.
I Think if first SCSI Controller fails, the vgxx in those disks go down. because I havent altenate path

If I can have alternate path, What can I do?

The question is:
Must I do "vgextend...." to have alternate path?

Thanks

[Steve Hamilton_2]

Jorge,

Yes - using 'vgextend' to bring the 2nd path to each of the drives into the existing VG will set up Alternate links, which will protect you against the loss of a single interface card or cable...

If you look at the VG using 'vgdisplay -v' after the 'vgextend' commands are complete, you will see the Alternate Links to each drive...

Steve Hamilton

[D Block 2]

Jorge, study this example script below..

run: setboot
does it show two paths to your boot (primary, and alternate) ?

The LVM commands below can help you understand the vgextend (NOTE: use your own c#t##d# numbers).

It sounds like you have (2BCC), when you have a portA active, then a portA (passive). only one controller is active, and the other will be used only when there is a Hardware Failure. the SCSI from the ds2300 has two cables to your L1000. only one SCSI cable does the work, then other is passive. (I'll assume you put an extra SCSI card into your L1000 for removing the Single-Point-Of-Failure -- one cable to the DS2300).

here's the script:

# cat mirror2.sh
# mirror'ed disk is: c3t15d0
##pvcreate -B /dev/rdsk/c3t15d0
##mkboot /dev/rdsk/c3t15d0 # r disk
##mkboot -a "hpux -lq (;0)/stand/vmunix" /dev/rdsk/c3t15d0
## lifcp /dev/dsk/c3t15d0:AUTO - # snd AUTO string to stdout
##vgextend /dev/vg00 /dev/dsk/c3t15d0

set -v

# for each lvol one through eight or nine
lvextend -m 1 /dev/vg00/lvol1 /dev/dsk/c3t15d0
lvextend -m 1 /dev/vg00/lvol2 /dev/dsk/c3t15d0
lvextend -m 1 /dev/vg00/lvol3 /dev/dsk/c3t15d0
lvextend -m 1 /dev/vg00/lvol4 /dev/dsk/c3t15d0
lvextend -m 1 /dev/vg00/lvol5 /dev/dsk/c3t15d0
lvextend -m 1 /dev/vg00/lvol6 /dev/dsk/c3t15d0
lvextend -m 1 /dev/vg00/lvol8 /dev/dsk/c3t15d0


lvlnboot -r /dev/vg00/lvol3 # Specifies the root logical volume
lvlnboot -s /dev/vg00/lvol2 # Specifies the swap logical volume
lvlnboot -d /dev/vg00/lvol2 # Specifies the dump logical volume
lvlnboot -b /dev/vg00/lvol1 # Specifies the boot logical volume
lvlnboot -R # Recovers any BDRA info
lvlnboot -v # Verify boot, root, swap and dump settings.