Re: multiple entries for the same file in the cache_swap_log file

> Andre Albsmeier writes:
>
> >OK, know I figured out how to interpret the lines from the
> >cache_swap_log file (the first entry contains the coded path of
> >the file containing the object), I wonder what these entries mean:
> >
> >00000020 3518b622 fffffffe 35182ed8 3401 http://static.hotbot.com/images/list6.gif
> >00000020 351a2042 fffffffe 34c257cb 682 http://www.freebsd.org/icons/back.gif
> >00000020 35376a75 fffffffe 34c27624 786 http://www.freebsd.org/gifs/littlelogo.gif
> >00000020 35376e2a fffffffe 3527e9fb 10381 http://www.korealink.com/public/files/buttons2.
> gif
> >
> >This would mean that the file resides in 00/02/00000020 when
> >16 and 256 directories are used.
> >
> >But of course there is only one URL for which the file is valid,
> >in this case it is the last:
> >
> >root@curry:/usr/local/squid/cache>ll 00/02/00000020
> >-rw-r--r-- 1 squid sys 10381 Apr 17 16:58 00/02/00000020
> >
> >So my question: How does it come that there are multpile
> >entries pointing to the same location?
>
> Swap file numbers get reused, for example, when old objects are removed or when
> an object gets updated with new content.
>
> When squid exits cleanly, the "storeWriteCleanLogs" function rewrites that file
> such that there are no duplications.
>
> Duane W.

I see. The reason I wanted to know that is because I would like to write
a small program that checks the logfile against the real contents of
the cache directories. If it finds any inconsistency it will delete
the log entry or the file, respectively.

Thanks for your explanation,

        -Andre
Received on Mon Apr 20 1998 - 21:50:00 MDT