TCP_SWAPFAIL issue

guys,

        i have squid setup in our network as shown in the diagram:

                    INTERNET
                       |
                       |
                 ------------ -----------
                 | parent | |child |
                 | squid |--------|squid-2 |
                 ------------ -----------
                      |
                      |
                 ------------
                 | child |
                 | squid-1 |
                 ------------

i have setup parent squid as the gateway of all squid.
child squid-1 and child squid-2 have the ff. entries:

        cache_peer proxy.gizmo.com parent 8888 0 no-query
        prefer_direct off

...
the problem is that some request from child squid-1 to parent squid have
TCP_SWAPFAIL but when i view the logs from the parent squid , it shows
TCP_MEM_HIT or TCP_HIT or TCP_IMS_HIT ...
the other child squid-2 doesnt have these TCP_SWAPFAIL in its access.log.
both child squids are configured from the parent proxy to have bandwidth
limits using squid's delay_pool feature. Is this how squid deals with
excess requests set by delay pool ?
please advice. thnx
Received on Fri May 05 2000 - 02:56:19 MDT