I guess my point is that it's sort of redundant right? If disk blocks underneath the volume have already had zeros written across them - then why do it again?
I'm trying to understand what are the real world implications. E.G Customer A didn't init a volume and had some bad things happen.
The initialization here is not redundant, as initializing the Logical Volume will be on file system level, while the RAID is on the block level. Our tests shows that initializing logical volumes are more stable.
I, too, am not getting it! Wht exactly is the advantage of zero'ing a iSCSI or FC volume? If I put toghether a new storage, there're probably only zeros on the drives anway, or any other pattern, that the manufacturer chose.
What, other than detecting any problematic sectors on the volume, could this be good for? And why only the initial amount and not the subsqeuent ones.
Seems that I could just create a 1 GB iSCSI Volume and have it zero'ed and afterwards I go ahead and grow it to the size I initially wanted?
You dont have to use this function and this was demanded by many users even EMC products will force you to use this function when creating an iSCSI LV. Initialization fills volumes with all zeros. It is for security reasons as uninitialized volumes may lead to data corruption. Also some drives may not be new and we would not be able to tell from the sw end, so this is a good reason to have it, again you dont have to use it.
the technical benefit still escapes me though… The only thing I could possibly think of being worth to be zeroed out are the areas where the FS would store its data, but the rest of the LV I wouldn't care about.
This also plays nicely with the statement that one wouldn't have to zero out the areas where the LV would have been grown to.