Peer Cache Statistics -> Are the statistics wrong or the code ???

I'v detected that using squid1.2beta20+patch1 the "Peer Cache Statistics"
reported by cachemgr.cgi, minutes after squid is started are about 120ms for
all peers, a reasonable value, but after some hours all the RTT values
become 0, and the cache.log reports incorrectly that some peers are
oscilating from DEAD to REVIVED state.

I'm i missing something or is some bug lost in the peer selection code ??

Bye & Thanks in advance for all the people that at least read this message
8))

[]---------------------------------------------------------------[]
  Pedro Ribeiro
  Online: http://www.cc.isel.pt/Pessoais/PedroRibeiro
  IRC(PTnet) Nick: PAntMaR
  e-Mail: Personal: pribeiro@cc.isel.pt
          Admin: admin@isel.pt
          Pager: pribeiro.pager@cc.isel.pt
          (PS: Short, NO MS Trash, plain text only, NO MIME Stuff)
  Tel: +351-1-8317128 / Fax: +351-1-8317171
[]---------------------------------------------------------------[]
Received on Wed May 20 1998 - 04:38:03 MDT