Hi, Due to a recent crash of my PC which subsequently resulted in corruption of my btrfs partition, I set out to research about file system reliablity in general and btrfs in specific. I came across the numurous accounts of corruption due to the use of btrfs with dm-crypt. I also found Chris Mason''s barrier test http://oss.oracle.com/~mason/barrier-test. Having tested ext4 FS with it, I could consistently corrupt it when on dm-crypt *with* barrier mount option on. As sad as it was finding about this issue, I decided to instead put a plain btrfs for my home directory and mount ecryptfs at the top of it for the time being. As for my backup btrfs partition which already has a lot of COW trees - about a year worth of backup, this isn''t a possibility. I therefore devised a different plan, however, I need to ensure a certain characteristic of btrfs before going ahead with it: Let''s assume we have a physical partition formatted as btrfs with no dm layers in play - so we know write-barriers work ok. We then put a large file (>500GB) on it called "apple". Using COW cp, we copy it to a new file called "orange". Now if we open orange and start writing randomly in it and while writing, pull the power plug, is it gauranteed that when we mount the fs next time, *apple* is intact - both in terms of metadata and actual data? Cheers, Mansour -- To unsubscribe from this list: send the line "unsubscribe linux-btrfs" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html