[squid-users] BUG 3556

[Stephen Borrill]

On 15/10/2020 16:02, Alex Rousskov wrote:
> On 10/15/20 10:06 AM, Stephen Borrill wrote:
>> I have also been pointed to your comment here:
>> https://bugs.squid-cache.org/show_bug.cgi?id=5069#c1
> 
> Aha, I thought there was another related bug but did not check. Glad you
> found it!
> 
>> I'm currently testing the COMM_ERROR -> NOMESSAGE patch as bug 5069
>> sounds identical.
> 
> Please update 5069 bug report if that code change works for you too.

Will do.

> It
> will give whoever volunteers the final solution more confidence that
> they are on the right path. The key here is whether the error has a
> transient nature and does not affect the listening socket ability to
> accept other transactions.

Yes, that's what I'm curious to know too. The problem seems to occur
once every few days and because the error will still be logged, it
should be easy to test whether operation continues afterwards.

-- 
Stephen