Assuming this WRITING state lasts for a while, it is possible that
network delays are to blame if your helper is using a network uplink and
does not read new helper requests while waiting for that network I/O.
> 2 squid
Assuming this WRITING state lasts for a while, it is possible that Squid
is to blame if your Squid is overloaded and cannot write to the helper
process (fast enough), accumulating more being-sent helper requests in
the helper queue.
> 3 helper
Assuming this WRITING state lasts for a while, it is possible that the
helper is to blame if your helper is overloaded (or stuck) and cannot
read requests from Squid fast enough.
The above general answers are not specific to your store_id helper. You
may narrow the list of suspects down by applying your knowledge about
your helper and Squid/OS statistics for its processes.
first thank you for the Helper_states page i could not find it
second i have 4 identical server 3 never have this issue
flag = BW
all server have very hi traffic and the timing max betwean 0.300 up to
just one server have the BW issue and cpu get hi as well one pc behind for
testing it was the ECC ram i never know it dose that and it raise squid
cpu% and fkp the I/O between squid and helper
thank you for your help