Hi, From: "Russell Coker" <[email protected]> > On Fri, 31 Oct 2014, "Peter Ross" <[email protected]> wrote: >> The article also mentions some speed issues especially in relation to >> databases. > > COW filesystems use different blocks on disk every time a file is written > to.
Yes. >> I would be interested to know what Oracle says to databases on ZFS on >> Solaris - and Btrfs on Linux systems (the later not supported by Oracle >> yet, I believe, the first I am not sure about) > > The same performance issues apply to BTRFS and ZFS. The significant > difference is that L2ARC and ZIL can mitigate such problems So Btrfs does not have this kind of features, I understand. Thanks for pointing this out. >> My gut feeling: Use Btrfs for "bread and butter" work and not if you need >> 101% reliability. With backups and mirrors and failovers (which may be >> in place anyway) you may be fine. > > If you want good reliability then you need backups and mirrors anyway. Yes, no doubt. But failover is a bit more tricky if you are dealing with instable nodes. It is harder to catch an error (which triggers a failover) if there are many ways to fail and some quite unexpected. For the purpose of explanation: A switched off computer is easier to detect than some data inconsistency which may "sleep" for a while before being detected. You are buggered if you have two nodes with different issues then. >> I just do not get my head around why a subvolumes is called subvolume if >> it is (according to the FAQ) comparable to a file system - you just can >> have many of them in a pool. > > A subvolume is represented inside BTRFS in much the same way a directory. You > can't mount one subvol without operating on the rest of the filesystem, so if > a filesystem is corrupted such that it can only be mounted RO then that > applies to all subvols. My tube may be glued to a wheel so you may have to replace the wheel in order to replace the tube. But I doubt that it makes my tube a wheel;-) Something that is represented in much the same ways as a directory does not look like a volume to me. Every "subvolume" has a root, it can be mounted everywhere on the system - as a file system does. It is organized on top of a block structure which abstracts from physical devices (usually referred to as logical volumes) and creates a directory structure on it - as a filesystem does. My way of thinking of things and analogies and my imagination feels slightly tricked by btrfs terminology;-) But I won't continue arguing. I have been walking through the mountains in a fog and I looked at a map and tried to make sense, until my friend rotated the map by 180 degrees and we realized how much we lost orientation... I guess you walked through the Btrfs forest a bit longer than me. Talking of forests - that's how I would call the whole "bunch of filesystems" directory structure on top of a pool because you can have many trees (filesystems) with their own roots in it (while a traditional block device has one only) Cheers Peter _______________________________________________ luv-main mailing list [email protected] http://lists.luv.asn.au/listinfo/luv-main
