> Hello All, > > Anybody know why I am getting "Rebuilding storage in > /usr/local/squid/var/cache (DIRTY)" in my cache.log when I restart squid? > I > read in an old message from Heinrich that it was because squid wasn't > shutdown properly. However, I have completely rebuilt squid > and still get the same messages. > > The problem is that when I redirect port 80 clients through squid now, > they just hang. There are no indications in any logs that the > clients are even making it to squid. However, if I disable the proxy > setting in a client, the connections are made immediately. > Squid has been working perfectly for months and now this DIRTY cache > message and clients hang. This is known behavior when Squid is overloaded with some task. A lot of work has been done since 2.5 to reduce this, but there are still some instances where it happens. What I believe is happening here is that Squid is busy processing the cache, and client requests are left queued up in the OS connection buffer. They will get handled, but is a slow race between the client connect timeout and whatever squid is doing. > > I did have a power failure that lasted longer than my UPS could support, > so the server died abruptly. This would account for the > improper shutdown that Heinrich mentioned, but now I have completely > rebuild the entire squid directory structure and squid. > > I am using squid-2.5.STABLE4! Please upgrade then :) > > Anybody know what is going on here and how I can correct it? > 'DIRTY' just means the swap.state was missing or unreadable (as happens after a bad shutdown or crash) and squid has to rebuild from a complete file scan of the cache_dir. If the swap.state exists, try stopping squid, deleting it, and then starting Squid. You will get one run of DIRTY, let it finish, then further restarts should work okay again. Amos
