Squid_ The Definitive Guide - Duane Wessels [200]
Default
store_objects_per_bucket 20
Example
store_objects_per_bucket 15
Related
store_avg_object_size
Name
client_db
Synopsis
Squid keeps a number of statistics for each cache client (IP address). You can view them by visiting the cache manager client_list page. The ClientInfo data structure is about 240 bytes on 32-bit systems and 300 bytes on 64-bit systems. If you have thousands of clients, this database can consume a significant amount of memory. You can disable this directive and free up that memory for other uses.
Syntax
client_db on|off
Default
client_db on
Example
client_db off
Name
netdb_low
Synopsis
The netdb database contains round-trip time and hop-count measurements derived from ICMP pings. This directive specifies the lower limit for the netdb replacement policy. In other words, when Squid is removing netdb entries, it stops when the total number reaches netdb_low.
Syntax
netdb_low N
Default
netdb_low 900
Example
netdb_low 9900
Related
netdb_high, query_icmp
Name
netdb_high
Synopsis
The netdb database contains round-trip time and hop-count measurements derived from ICMP pings. This directive specifies an upper limit on the number entries in the database. When Squid finds more than netdb_high entries, it removes least-recently used networks until the size reaches netdb_low.
Syntax
netdb_high N
Default
netdb_high 1000
Example
netdb_high 10000
Related
netdb_low, query_icmp
Name
netdb_ping_period
Synopsis
This directive specifies how long Squid must wait between sending consecutive ICMP pings to the same /24 network. The interval is relatively long so that Squid's ICMP traffic doesn't upset server administrators.
Syntax
netdb_ping_period time-specification
Default
netdb_ping_period 5 min
Example
netdb_ping_period 3 min
Related
pinger_program, query_icmp
Name
query_icmp
Synopsis
Enabling this directive instructs Squid to ask its neighbors for their ICMP measurements, which are included in ICP/HTCP replies. This, essentially, populates your netdb database with your neighbors' ICMP measurements. The bulk "netdb exchange" is another way to receive those measurements (see Section 10.5).
Squid uses the neighbors' netdb measurements when making forwarding decisions. If one of the parents is closer to the origin server, Squid forwards the request there and marks it with CLOSEST_PARENT_MISS in access.log.
Syntax
query_icmp on|off
Default
query_icmp off
Example
query_icmp on
Related
pinger_program, netdb_ping_period
Name
test_reachability
Synopsis
When you enable this directive, Squid looks at its netdb database while processing ICP queries. If Squid normally returns ICP_MISS, but the origin server isn't in the database or doesn't respond to ICMP pings, it returns ICP_MISS_NOFETCH instead. The ICP_MISS_NOFETCH reply signals the neighbor cache that Squid might not be able to communicate with the origin server.
Syntax
test_reachability on|off
Default
test_reachability off
Example
test_reachability on
Related
pinger_program, query_icmp, netdb_ping_period
Name
buffered_logs
Synopsis
While this directive used to affect multiple log files, it now only applies to cache.log. Squid uses the stdio library for cache.log. If this directive is enabled, Squid calls fflush( ) after every write. This allows you to see log file entries as they are written. You might want to disable buffered_logs if you are debugging Squid in a way that creates a large number of cache.log entries.
Syntax
buffered_logs on|off
Default
buffered_logs off
Example
buffered_logs on
Related
cache_log
Name
reload_into_ims
Synopsis
If you enable this directive, Squid adds an If-Modified-Since header to requests that contain a no-cache directive. This is a global version of the reload-into-ims option for the refresh_pattern directive (see Section 7.7).
* * *
Warning
Altering the client's request in this manner is a violation of HTTP.
* * *
Syntax
reload_into_ims