Defragging shared array (Microsoft Clustering)

Oh man! Ad-blocking software has been detected! :'(

This website is run by the community, for the community... and it needs advertisements in order to keep running. Blocking our ads means your killing our stats!
Please disable your ad-block, or become a premium member to hide all advertisements and this notice.

Hi guys

We have a two node cluster (consisting of two HP Proliant servers) that are both connected to a shared array via SCSI.

To be honest, the drives could do with defragging as they haven't been for a while now. As both servers can see the same drives in the shared array, do I only need to log into one of the servers and clean/defrag the drives?

It's a bit of a silly question, but my boss left me a message asking me to log onto both physical servers and clean/defrag the shared drives. This doesn't make sense to me as it's effectively doing the same thing twice.

We have a total of six virtual servers setup on the two node cluster - so I'm not quite sure how to go about defragging.

Any help would be appreciated.

Thanks!

 

Oh man! Ad-blocking software has been detected! :'(

This website is run by the community, for the community... and it needs advertisements in order to keep running. Blocking our ads means your killing our stats!
Please disable your ad-block, or become a premium member to hide all advertisements and this notice.

To be honest, I haven't defragged a server in forever. Does anyone defrag anymore, really? In recent years, I've never noticed a slowdown by not doing it... my servers were always lightning fast without it.

 

I think more information would be helpful like what's your servers load balancing set to if any or Raid 1 or 5 etc.

To carry out the above if its Raid 1 or Raid 5 setup, just right click on the required disk and do a defrag considering you have fault tolerance in place.

 

With modern drives I've never felt the need to defrag and the thought of the possible data loss sends shivers down my spine .... course with a fully implemented Disaster Recovery plan it's not a problem

 

I suppose a defrag won't do them any harm!

onoski - Apologies if I've misunderstood your question - the cluster isn't load balanced, we have print and Virus Scanner services on one physical server and Backup, Email and file shares on the other server. So the cluster offers redundancy in case one of the servers fail.

If one server fails, they all failover to the working server. We have another two servers for domain controllers which are load balanced.

Each service has it's own virtual server name, IP address and storage/disk space on the shared array.

I hope I've explained this ok!

Some of the drives accessible from node1 and node2 seem to be the same, so I wasn't sure if I had to defrag from both nodes.

 

Having compared what drives are available from both nodes - it was my mistake really as the Quorum patition/drive is the only one that is accessble from both servers. But you can only access the contents from one server so I suppose I should have checked this before posting!

Been reading some articles about clusters - it advises to defrag by logging in with the virtual server name (rather than the actual nodes name) in case of failure - the defrag will just continue when the service fails over to another server - which obviously makes sense!

 

I, personally, would not recommend defragging a RAID 5 volume unless the vendor of the RAID controller it sits on recommends it. There are many reasons for this - most of them far too boring to go into - but it boils down to the way Windoze 'sees' a partition. Because it always assumes that a partition is on a single physical drive it will attempt to arrange data based on that - ignoring the fact that the underlying physical disk structure may be spread over dozens of disks. Consider, for instance, your stripe width. Say its set to 1024K and you have an 8192K file. The stripe will spread that across eight disks - which Windoze itself knows sod all about. I would argue hwere that defragging carries more risk of encountering data corruption (though still small) because you are putting more grunt through the disks, thus potentially causing one (or, catastrophically, more than one) disk to fail.

I'd ask why you're being asked to defrag them anyway? Is it because performance is poor? Perhaps you've got a pagefile sat on a RAID5 array (never a good thing - they should always be on RAID 0 or non-RAID arrays)? I agree with earlier posters - I can count on the fingers of one knee the number of times I've seen performance improve even infinitesimally after a defrag in the past six years.

 

Thanks for the replies and advice - this is why I love this forum so much!