BladeSystem Forums have moved here
To make BladeSystem information easier to find, we have moved the BladeSystem forums here, to Servers and Operating Systems.
Showing results for 
Search instead for 
Do you mean 

Planning for Disk Storage Outage

Honored Contributor

Planning for Disk Storage Outage

Need to know is it possible to automate or not?

Environment: HP-UX serviceguard cluster A.11.16 version with two node and each node at one data center connected to IBM shark storage. Storage has mirror copy in another data center.

As of now, we had Storage frame outage at one data center so we did switch to mirror copy disk sets (at another data center) manually with cluster downtime. Want to automate this process to detect Storage Frame outage and switched to mirror copy without user intervention.

1) Do I need any extra software?
2) Can I do it with scripts?

Exalted Contributor

Re: Planning for Disk Storage Outage


1) I think the highly reliable way to go is the Metro Cluster product from hp.

2) Yes but It will take significant testing to make sure its reliable.

Steven E Protter
Owner of ISN Corporation
Honored Contributor

Re: Planning for Disk Storage Outage

Research the metro cluster version of the MC/SG software.

Honored Contributor

Re: Planning for Disk Storage Outage

Both data centers are within 10 mile radius. Both nodes are in same subnet due to fibre connectivity between two data centers. Means, as if they are in one data center.

Metro cluster design didn't give me much info for this kind of question.

Looking for automation of mirror disk availablity when primary Storage Frame failure happens. Remember that Nodes are not down.

any other thoughts ??
Acclaimed Contributor

Re: Planning for Disk Storage Outage

Means, as if they are in one data center.

IF this can be taken as a true statement then it's rather simple to do what you want to do and MC/SG really won't even come into play. You could setup VG's using LUN's on each array as PV's. The LUN's from each array would constitute a separate PVG. You then simply setup LVOL mirroring and the host computer will continue happily if either
array fails. Each physical volume should have primary and alternate paths so that a Fibre failure will simply cause an alternate path activation rather than being considered an array failure. While I have explained this in LVM/MirrorUX terms, it applies equally well to VxVM.

If it ain't broke, I can fix that.
Honored Contributor

Re: Planning for Disk Storage Outage


Don't want to do LVM mirroring since Storage disks are mirrored at hardware level. Both nodes are having access to one Storage frame only when cluster is active. Nodes can't see my mirror disks from secondary storage at DR center.

My question is: If I make second data center storage available to nodes as read-only and when Primary Data center failure happens, script / software should activate disks and mount them for Cluster ??
Trusted Contributor

Re: Planning for Disk Storage Outage

Would you be able to setup an montior process as part of your package. This could keep an eye on storage frames, and should a problem occur fail the package to another node ?

Honored Contributor

Re: Planning for Disk Storage Outage

yes I can. But would you please explain how do I monitor frame outage and switch packages to DR site with activating Mirror copy from DR Storage?

Looking at DR planning for Metro clusters and which supports cluster config for EVA, XP and EMC frames. Don't know whether I can take same steps and implement for Shark Storage?