Hello Squid users,
My cache log contains these two entries:
2010/01/04 14:44:56| storeSwapOutFileClosed: SwapOut complete:
'http://www.somesite.com/somepage' to 7, 003339B9
2010/01/04 14:44:56| storeDirSwapLog: SWAP_LOG_ADD
ECA20C9F21F6770179851C89C2386046 7 003339B9
I tried to locate the file by name "003339B9" in each of my
cache_dirs, but all those filenames that match don't point to the URL
mentioned in the log above. In fact those files have been create like
a month ago. Isn't Squid supposed to save the file by the name given
in SWAP_LOG_ADD entry?
The above request was a TCP_MISS, and hence the swapOut. Few minutes
later, I try to fetch the same URL, and I get a TCP_MEM_HIT with an
entry like this in the cache_log:
2010/01/04 14:45:41| storeClientCopy3: Copying from memory
A day later when I fetch the same URL, it gets reported as a TCP_MISS.
I can't understand where did Squid save the cached copy. Going by the
filename in SWAP_LOG_ADD, I don't find anything relevant. As long as
the content was in memory, Squid considered it a hit. Once it got
flushed out of the memory, Squid couldn't possibly locate the cached
file in its cache_dir like I couldn't. Or may be the cache
replacement/maintainSwapSpace process removed the cached content
wherever it was saved. Any explanation?
Also, the refresh_pattern for my Squid is configured like 21600 100%
21600. So files stay fresh for 15 days(21600=15*24*60). And squid is
never restarted. So, is the store_table always up-to-date with the
files
stored on the disk?
TIA.
Received on Wed Jan 06 2010 - 02:17:41 MST
This archive was generated by hypermail 2.2.0 : Wed Jan 06 2010 - 12:00:02 MST