> the TCP_MISS_ABORTED seems down request and i dont know whats the
> wrong there
Assuming the second column in the sanitized access.log output below is
transaction response time, it looks like the client aborts (i.e. closes
the connection) if it receives no response from Squid for ~38 seconds. I
do not know why the server did not respond for so long or why Squid did
not forward the response to the client for so long. You may be able to
distinguish these two cases using a packet capture.
What logformat do you use? Did you reorder these records or were they
logged by Squid in the decreasing timestamp order?
The requests in each of the two groups above look concurrent to me. Do
you explicitly turn collapsed_forwarding on in your squid.conf? If you
do, does turning collapsed forwarding off solve this problem?
> as long as i can ping it from an iPV6 address
This is probably not an IP connectivity problem but a TCP- or HTTP-level
problem. You may want to try curl or wget to reproduce it without Squid.