[rhelv6-beta-list] My first experiences with RHEL6 beta

[Lamar Owen]

On Wednesday, June 16, 2010 10:10:44 am R P Herrold wrote:
> - The older methods are almost self documenting on a single 
> man page; in managing adding a new PV recently, I was toggling 
> between several

If you like a GUI, use palimpsest.  Really, through ssh X tunneling if needed.  It's looking really good, even if a little rough around the edges.

Otherwise, 'man lvm' gives you the core of what you need to know.  And it's not hard at all, really, not compared to some other things.  And the flexibility in an enterprise environment allows you to manage storage in much the same way as you'd manage LUN's through Navisphere (for EMC gear), but with non-SAN disks.

> - The older Linux config method of hanging a stated mountpoint 
> on a stated partition of a stated drive is 'sed and awk' 
> friendly for scripted management tools; the LVM method 
> requires deeper knowledge of what is happening and is less 
> well scriptable

But what happens when disk detect order is nondeterministic?  I've fought and fought boxes with multiple controllers where /dev/sda on one boot would be /dev/sdc on another thanks to parallel module loading and non-linear init setups.  And with hotplugging and hotunplugging on enterprise hardware being more the norm than the exception, even on boxes without hardware RAID you can do hot data migrations with LVM on mounted filesystems that you simply cannot do with the older tools.

> - The older methods are more robust and less fragile. 

Perhaps.  At least until I add a disk on the third SCSI chain of a system with seven SCSI controllers installed (been there, done that, on our Sun E6500 a while back, running older Aurora SPARC Linux 2.x; definitely enterprise hardware).  The solution in fstab-land is UUID or label based mounting; LVM just gives better management tools.  And with the 20 plus minutes that an E6500 can take to POST, rebooting to do anything but kernel upgrades is really a non-starter.  LVM really lets you dynamically administer your storage with little to no downtime, in a fairly simple manner that is getting more stable all the time.

But it does take more effort to recover, yes, I can see that.