RAID stripe size

[Andyrh]

I am about to rebuild my RAID group. Is there a stripe size that would be better?
Does ZM write events in any particular chunks?
For instance, MS SQL reads and writes in 64k blocks, so aligning the file system and stripe size to 64k makes a difference.

[iconnor]

No we don't. We write out in whatever size the packets are, which generally are much smaller than 64k. We rely on the OS to cache and do the right thing.

Optimising the db is def a good idea, but I wouldn't worry too much about video storage. It would actually make sense to write to a ramdisk first, then copy to disk. Because we write the data out, then go back to write the table of contents at the beginning.

ONe of the reasons I recommend if you are going to use NAS is to write local, then move to NAS in background.

[Andyrh]

I suspected there was not a "good" size, but asking is better than guessing.