[squid-users] All Adaptation ICAPs go down at the same time
Alex Rousskov
rousskov at measurement-factory.com
Wed May 19 21:53:28 UTC 2021
On 5/19/21 5:31 PM, roie rachamim wrote:
> 2021/05/12 12:27:24.209| 93,5| AsyncJob.cc(139) callEnd:
> AsyncJob::start() ends job [/ job31640]
To me, this looks like bug 4528:
https://bugs.squid-cache.org/show_bug.cgi?id=4528
That bug is being fixed in PR 795:
https://github.com/squid-cache/squid/pull/795
You might be able to patch your Squid using the corresponding patch:
https://github.com/squid-cache/squid/pull/795.diff
Testing feedback is very weclomed.
HTH,
Alex.
> On Mon, Apr 19, 2021 at 12:07 PM Eliezer Croitoru wrote:
>
> Hey Roie,____
>
> __ __
>
> From the output I assume it’s a dns resolution issue.____
>
> In the past I remember that Docker was updating the hosts file with
> the relevant names but it’s not working the same way now.____
>
> Currently Docker is using a local network dns service which is being
> accessed via 127.0.0.53.____
>
> From I remember Squid is resolving the icap service name only at
> startup or reload.____
>
> Lately Alex published a testable patch that might fix specific
> issues with icap services which are resolved by dns. ( sorry I don’t
> remember the bug report)____
>
> I assume you can try to test this patch first.____
>
> If these services are static to some degree you might be able to
> create a script that updates the hosts file and reload squid on each
> change.____
>
> When using the hosts file it’s possible that some issues will
> disappear.____
>
>
> There is also another possibility which is a malformed ICAP response
> or wrong sessions handling which cause this issue.____
>
> You might be able to use tcpdump from either the host or the
> container side to capture traffic when these goes down.____
>
> Depends on your preference of debug level you might even be able to
> debug specific debug_options like for ICAP services
> and/or requests to the degree you might be able to see what happens
> on the basic level of the ICAP encapsulation.____
>
> If you really need help with a diagnosis and a solution you might be
> able to use Alex and the measurement factory.
>
> ____
>
> All The Bests,____
>
> Eliezer____
>
> __ __
>
> *From:* squid-users <squid-users-bounces at lists.squid-cache.org
> <mailto:squid-users-bounces at lists.squid-cache.org>> *On Behalf Of
> *roie rachamim
> *Sent:* Monday, April 12, 2021 12:54 PM
> *To:* squid-users at lists.squid-cache.org
> <mailto:squid-users at lists.squid-cache.org>
> *Subject:* [squid-users] All Adaptation ICAPs go down at the same
> time____
>
> __ __
>
> Hi,____
>
> __ __
>
> Our setup includes squid that runs in docker container with several
> ICAP servers in additional containers.____
>
> From time to time we see in cache.log the following messages:
> 2021/04/12 00:22:39| optional ICAP service is down after an options
> fetch failure: icap://icap1.proxy:14590/censor [down,!opt]
> 2021/04/12 00:22:39| optional ICAP service is down after an options
> fetch failure: icap://icap2.proxy:1344/request [down,!opt]
> 2021/04/12 00:22:39| optional ICAP service is down after an options
> fetch failure: icap://icap3.proxy:14590/response [down,!opt]
>
> 2021/04/12 06:10:45| optional ICAP service is down after an options
> fetch failure: icap://icap1.proxy:14590/censor [down,!opt]
> 2021/04/12 06:10:45| optional ICAP service is down after an options
> fetch failure: icap://icap2.proxy:1344/request [down,!opt]
> 2021/04/12 06:10:45| optional ICAP service is down after an options
> fetch failure: icap://icap3.proxy:14590/response [down,!opt]____
>
> __ __
>
> We're trying to understand why it happens to all ICAPs at once. This
> happens in 4.14 and in 5.0.4____
>
> Any thoughts about what might cause this ?____
>
> Many Thanks,____
>
> Roie____
>
>
> _______________________________________________
> squid-users mailing list
> squid-users at lists.squid-cache.org
> http://lists.squid-cache.org/listinfo/squid-users
>
More information about the squid-users
mailing list