On Thu, 26 Jan 2012 18:47:10 +1300 Amos Jeffries wrote: > > I assumed that part of the point of having a journal was that > > reading it in sequentially would lru order the queue when squid > > starts up. > > You are right. I checked the code and it uses the policy list to dump > the clean journal. > > I was referring to deleting swap.state when squid isn't running, in > > which case it's recreated from the cache files. > > If you do this erase you will loose the LRU atime details in the > journal. Right, that's what I was saying when I said "I'm guessing that the rebuild had already been done automatically". My theory is that that an unclean shutdown, or something, caused swap.state to be lost or corrupted, forcing squid to read in the objects in the cache. Presumably the lru queue would then get filled in hexadecimal path/filename order, which would account for the bogus LRU reference ages. The reason that I tried a manual slow rebuild first is that it seemed very likely to be lru corruption and I though there was a small chance that squid might do a sort on a slow rebuild. The fact that the reference age didn't change at all suggested that squid had already done its own slow rebuild on a previous restart. If the files are read in hexadecimal order then it should be fairly straightforward to repair the lru information by reordering the files in the cache by mtime and having squid reread them.
