On Mon, 16 May 2011 13:29:35 +0200, Bruno Cornec <[email protected]> wrote: > Exactly. And typically I'd like to avoid backuping /var, except maybe > /var/spool. Currently what I *have* to backup is crontab files > (/var/spool/cron) > , /var/spool/mail, /var/spool/news, /var/lib/mysql, and /var/www where > most of the rest is useless from a recovery perspective. Absolutely! That's also why I suggested to use /srv in many cases e.g. for databases.
> Moving useful data from /var/spool to /srv/spool or something similar > would help as we could then just backup the full /srv and avoid /var > content at all. I still would keep this temporary/volatile data in /var ... even though this makes backups more "complex". Cause typically all this spooling stuff should be there only for a short period of time, until it's further processed (e.g. handed over to the printer or to a pop3/imap server). > For me the real aspect is data separation, filesystem coherency > (possibility to have one broken such as /var due to logs filling it, > random temp data written to it, ...) where the other one is more under > control (backup being part of it). Agreed. Cheers, Chris. _______________________________________________ fhs-discuss mailing list [email protected] https://lists.linux-foundation.org/mailman/listinfo/fhs-discuss
