Somehow I missed the original email... On 01/08/15 21:32, Daniel Parks wrote: > I'm afraid I haven't had a chance to test this with tarsnap, but typically > the way to do this is with mkfifo, which creates a named pipe (a FIFO: First > In First Out). An example from memory: > > $ mkfifo mysqldump.sql > $ mysqldump mydb > db.sql & > $ tarsnap -cf myarchive db.sql > > Tarsnap should read the output of mysqldump and store it as db.sql. The > db.sql named pipe on disk doesn't contain the data from mysqldump; if you try > to read it again it will hang waiting for input from another process.
This doesn't work; tarsnap doesn't archive fifos. >> On Jan 3, 2015, at 9:27 AM, Philipp Riegger <[email protected]> wrote: >> I am looking for a backup solution for mysql databases and subversion >> repositories. I think it would be more interesting to backup the dumps >> from "svnadmin dump" and the corresponding mysql command. >> >> Saving the dumps each time wastes a lot of disk space (and produces a >> lot of unnecessary disk IO). Is there a way to save streams (or data >> from stdin)? Since data from stdin has no timestamp and no user/group >> information, maybe there is a way to create a fake tar file from stdin >> and pipe that to tarsnap? I did some searching on the internet and read >> the man page. This does not seem to be possible at the moment. Is that >> correct? You can't pipe data into tarsnap. The reason for this is that tarsnap uses the tar format internally, and so it needs to know the size of a file before it starts to read the file data. You can pipe an archive into tarsnap (e.g., `... | tarsnap -c -f foo @-`) but this doesn't really solve the problem -- you need to know the amount of data you're going to stream in before you start streaming the data, in order to create the necessary archive header for tarsnap. -- Colin Percival Security Officer Emeritus, FreeBSD | The power to serve Founder, Tarsnap | www.tarsnap.com | Online backups for the truly paranoid
