the FAQ talks about how to limit per-directory size by the function
of <N> filesystems being *cache_swap* / <N>
4.13 Is it possible to control how big each cache_dir is?
With Squid-1.1 it is NOT possible. Each cache_dir is
assumed to be the same size. The cache_swap setting
defines the size of all cache_dir's taken together. If
you have N cache_dir's then each one will hold cache_swap
� N Megabytes.
but:
eeyore # cd lib/
eeyore # grep -i cache_swap *.[ch]
eeyore # cd ../src/
eeyore # grep -i cache_swap *.[ch]
cache_cf.c: fatal("cache_swap is lower than cache_mem");
cf_parser.c: default_line("cache_swap_low 90");
cf_parser.c: default_line("cache_swap_high 95");
cf_parser.c: /* No default for cache_swap_log */
cf_parser.c: } else if (!strcmp(token, "cache_swap_low")) {
cf_parser.c: } else if (!strcmp(token, "cache_swap_high")) {
cf_parser.c: } else if (!strcmp(token, "cache_swap_log")) {
cf_parser.c: dump_int(entry, "cache_swap_low", Config.Swap.lowWaterMark);
cf_parser.c: dump_int(entry, "cache_swap_high", Config.Swap.highWaterMark);
cf_parser.c: dump_string(entry, "cache_swap_log", Config.Log.swap);
eeyore #
I.E. there is no directive named cache_swap, nor from what I can see is
there ANY directive which limits the maximum on-disk size. Its assumed to
grow to the low/high water marks of the complete FS size..
I can't from simple greps see how
globals.h:extern int store_swap_size; /* 0 *
ever gets set by cmdline or config file value either. Since the code
has
store.c: Config.Swap.maxSize = store_swap_size;
it appears that until store_swap_size gets set (from a stat() on the device?)
its not defined..
I'm being thick. I know it. So how do you constrain Squid to use a defined
limit on available filesystem space?
-George
Received on Wed Feb 04 1998 - 16:18:31 MST
This archive was generated by hypermail pre-2.1.9 : Tue Dec 09 2003 - 16:38:47 MST