So the basic setup here is that your origin will have to handle the
manifest processing and delivery of media offsets to ATS.
>From that point, ATS can cache the origin responses and deliver those
cached responses to the end user.
So your flow would be something like:
end user <-> ATS <-> MSS Or
Does ATS support sending a TCP reset for a denied request instead of
sending back a layer 7(HTTP) response ?
There are some end user requests that if they meet a certain condition
that I would rather just RST the connection instead of responding with
a 4xx /5xx response code.
On a side note. I put together an internal POC that instantiates a
high performance ATS docker container within a multi-tenant
environment.
I used cgroups to allocate cpu/memory resources from within the same
NUMA node.
I also used a docker plugin to pass through SRIOV VFs, also within the
same NU
Can you post the complete records.config from your child cache ?
If I had to guess, it looks like heurstics or fuzzytimer may be set too low ???
But if you provide a company sanitized version of your records.config
that should give us a better understanding of your setup.
On Wed, Jan 3, 2018
ttps://gist.github.com/poblahblahblah/a1abdad3d64ebdfba003c75e50c9fe00
>
> Thanks for looking!
>
> On Sun, Jan 7, 2018 at 8:13 PM, Jeremy Payne wrote:
>> Can you post the complete records.config from your child cache ?
>> If I had to guess, it looks like heurstics or fuzzytimer may b
What is the current status of adding the CARP plugin as an 'official'
ATS plugin ? Or at least
classified as an 'experimental' plugin ?
Or has CARP been scrapped as the community trends towards upfront
content routing combined with edge-to-parent consistent hashing ?
Jeremy
I don't cover entry points and building the container image. That's
the easy part anyway.
What you'll find here are steps on passing resources directly to an
instantiated container.
http://mentalsandbox2020.blogspot.com/2018/01/high-performance-ats-container-in-multi.html
On Mon, Jan 8, 2018 at
I've been thinking about this, but you can use current parent
functionality to perform local node(edge) object lookups.
It's a hack and requires careful configuration of your remap and
parent configuration so to not end up in a cache loop
or end up storing duplicate copies.. but it is a doable work
btw... i'll add a lesson learned post to my wiki..
give me a couple days to clean up lab details.. :-)
On Wed, Jan 17, 2018 at 1:36 PM, Jeremy Payne wrote:
> I've been thinking about this, but you can use current parent
> functionality to perform local node(edge) object lookup
Is there an over-rideable configuration or maybe an API function I can
use to disable write locks for
a given cache-key/txn ?
Basically after the cache lookup is complete, I want to parse request
headers and if header values match a certain string/condition, I want
to disable the
write lock. I'm
Is it safe to conclude that in terms of request routing, that this CVE
only applies to proxies in forward proxy mode ? Or rather forward
proxies that parse the host header to determine next hop ?
In reverse proxy mode, where remap rules are explicitly defined, then
a request either matches a remap
I've been running daily load tests for almost the last 6+ months.
6000 simulated users, which is 12k connections at around 17-18gbps.
Before I started playing with the collapse_forwarding plugin, I would see
requests 'leaking' to the origin, when those requests should have been
served by cache.
Now
One thought is to turn on debug with parent routing enable then compare
against debug with parent routing disabled.
Do you see any extra or missing steps while cache is being looked up(read)
or when the open write lock is created ?
On Thu, Mar 8, 2018 at 1:25 PM, Dunkin, Nick wrote:
> HI Su
Perfect use case for the lua plugin.
On Thu, Mar 22, 2018 at 1:49 AM, Vasanth Mathivanan
wrote:
> HI,
>
> We are working HLS live Concept in Traffic server ,when the client request
> made hit to get the m3u8 file then ts files …My question how get content
> before request to stored two or three
? In the document so many
> configuration which one can I use ?
>
>
>
> Sent from Mail for Windows 10
>
>
>
> From: Jeremy Payne
> Sent: Thursday, March 22, 2018 8:20 PM
> To: users@trafficserver.apache.org
> Subject: Re: Need Some Adjustments in Live Scenario
last
> three .ts file before request to stored in cache ?
>
>
>
> Thanks & Regards
>
> Vasanth M
>
>
>
> Sent from Mail for Windows 10
>
>
>
> From: Jeremy Payne
> Sent: Thursday, March 22, 2018 10:04 PM
>
>
> To: users@trafficserver.ap
Context:
ATS server terminating multiple secure sites
Question:
Is there an out-the-box configuration that allows me to
require client certificates if the client hello/handshake matches a
specific SNI value?
Something like:
+++
if
SNI matches foor.bar
then
proxy.config.ssl.client.certificati
>{ fqdn="example.com", verify_client=MODERATE },
>{ fqdn="*.yahoo.com", verify_client=STRICT }}
>
>
> Syeda Persia Aziz
> Software Developer
> Yahoo! Inc.
> Champaign, Illinois
>
>
> On Tuesday, March 27, 2018, 2:08:18 PM CDT, Jeremy Payne &l
I think the issue is the child is sending a https request to the parent.
Does the parent support https on port 3128 ?
On Wed, Mar 28, 2018 at 10:45 AM, Veiko Kukk wrote:
> Hi,
>
> I'm trying to get ATS 7.1.2 working with single parent and failover to
> origin.
> "clients" make request against ATS
- / - NONE/- text/htm
>
> Veiko
>
>
> 2018-03-28 19:01 GMT+03:00 Jeremy Payne :
>>
>> I think the issue is the child is sending a https request to the parent.
>> Does the parent support https on port 3128 ?
>>
>> On Wed, Mar 28, 2018 at 10:45 AM, Veiko Ku
In looking at 7.1.3 and master documentation I notice that
proxy.config.hostdb.strict_round_robin
is only reloadable not over-rideable.
Are there plans to make this over-ridable or is there another ATS
parameter that provides same functionality but on a per remap basis.
Thanks!
ure. We're hoping to do better in the long run as we
> update parent selection and related mechanisms. Do you really want that to
> be overridable per transaction or per host?
>
> On Tue, May 1, 2018 at 3:44 PM, Jeremy Payne wrote:
>>
>> In looking at 7.1.3 and ma
Per end user request, does ATS generate a unique ID I can also use as
a transaction id?
I am thinking of creating an ID(hash) based on URL+some-random-string,
but want to make sure something isnt generated internally(and exposed
via some API), before starting this work.
The ID will then be added t
t; On Mon, Jun 11, 2018 at 3:45 PM, Jeremy Payne wrote:
>>
>> Per end user request, does ATS generate a unique ID I can also use as
>> a transaction id?
>> I am thinking of creating an ID(hash) based on URL+some-random-string,
>> but want to make sure something isnt g
in testing 713, i also notice the 303 sent back to the end user. this
is much different than what's seen with 621. at least with 621 a 5xx
reponse is sent after all configured retries have been exhausted.
that said.. whats the current status of this plugin? deprecated? no
longer supported ?
is i
tiple instantaneous hits for the same
> content. So this actually comes up fairly often for linear video since you
> have multiple clients that can all be requesting the same video chunk at the
> same time
>
> On 2018/06/19 12:24:08, Jeremy Payne wrote:
>> in testing 713, i
to end up attempting another write and if the timing is just
> right the first write may finish right as another request failed to read and
> attempted to do another write, which will result in an origin hit. But thats
> something to figure out once Ive worked out the core issue
>
If you're open, you'll find what you're looking for in the lua plugin.
On Mon, Jun 25, 2018 at 12:02 PM, Chillin Gong wrote:
> Hi,
>
> I am a new user of traffic server. Recently I am trying to migrate a product
> from ATS5 to latest ATS7. In ATS5, we use the cache feature, and use the
> cacheu
hanks for the reply! Do you mean there are some lua scripts that implement
> the cache key feature, and they can be hooked into traffic server with lua
> plugin? Do you have a sample lua script for it?
>
>
> Regards,
> Chunlin
>
> On Mon, Jun 25, 2018 at 3:56 PM, Jeremy P
does ats support h2c or still a work in progress ?
thanks!
As always, thanks for your response.
Jeremy
On Thu, Aug 9, 2018 at 6:22 PM Leif Hedstrom wrote:
>
>
>
> > On Aug 9, 2018, at 5:03 PM, Jeremy Payne wrote:
> >
> > does ats support h2c or still a work in progress ?
>
>
>
> It does not, and there are no pla
Was collapsed_connection deprecated due to some technical flaw or
incompatibility ? or just from a lack of development ownership ?
I understand that collapsed_forwarding attempts to solve the same
issue, but collapsed_connection seems to be more forgiving than
collapsed_forwarding(at least the open
Context:
Openssl 102k
ATS 714
I notice that at times a client will send a TLS 1.2 close-notify,
immediately followed by a FIN-ACK. Which seems to be following spec.
"It is not required for the initiator of the close to wait for the
responding close_notify alert before
closing the read side of
Anyone have luck using tcpinfo and the two 'close' hook points?
ssn_close
txn_close
If I use either of these hook points, the plugin throws an error in
traffic.out(debug) and doesn't log into the specified tcpinfo log
file.
ex - tcpinfo.so --log-level=2 --hooks=txn_close
Hook points *_start an
Maybe use the below plugin to exclude the FQDN as the part of the
cache key or generate the cache key based on the path+file ?
https://docs.trafficserver.apache.org/en/7.1.x/admin-guide/plugins/cachekey.en.html
On Wed, Oct 10, 2018 at 4:06 PM Jin wrote:
>
> I have a backend server using DNS roun
What version of ATS and are you using the collapsed forwarding plugin?
On Mon, Oct 15, 2018 at 2:28 AM Vasanth Mathivanan
wrote:
>
> Hi,
>
>
>
> We have problem on trafficserver error log 502 response at same time check
> the “custom_logs” get that particular request “pssc=303” , then again get t
> ___Vasanth
>
>
>
> Sent from Mail for Windows 10
>
>
>
> From: Jeremy Payne
> Sent: Monday, October 15, 2018 6:40 PM
> To: users@trafficserver.apache.org
> Subject: Re: TrafficServer 502 Error
>
>
>
> What version of ATS and are you using the collapsed
re: 502 reponse
Is your origin not available during this time or are you hitting the
connection(aka first-byte) timeout?
proxy.config.http.connect_attempts_timeout
Also.. I believe you are showing the error log entries.. do you see
corresponding 303s for the same timestamp in the access log ?
The same functionality is in parent selection.
Try both and compare.
On Fri, Oct 26, 2018, 11:29 AM Vasanth Mathivanan <
vasant...@evolutiondigital.com> wrote:
> Hi,
>
>
>
> getting every response 80 and 81 , but our origin server having only 81
> port and give it also parameters , so its came a
gt;
>
> “dest_domain=x.x.x.x. port=81 method=get parent="10.137.10.174:8087|1.0;
> 10.137.10.173:8003|2.0" round_robin=consistent_hash”
>
>
>
>
>
> __Vasanth
>
>
>
> Sent from Mail for Windows 10
>
>
>
> From: Jeremy Payne
> Sent: Friday,
Maybe I'm missing something here..
If the main purpose here is to issue a 'Connection:close'
why not just disable client facing http keep-alive ?
traffic_ctl config set proxy.config.http.keep_alive_enabled_in 0
After the param has been activated, requests on existing connections
will receive a C
Can you show the MPD request at the origin layer as well.. Again for
the same title seen in your email.
If I had to bet it looks like there is something amiss with the MPD
cached at the ATS layer.
BTW, the Roku does send a CLIENT_ABORT(TCP RST) as part of some
discovery mechanism it performs.
But
it appears you have a plugin and/or cache-key conflict.
meaning, you are sending the full object back when the client is
expecting a specific range.
so the roku may be sensitive to that..
are you removing the 'range' header between ATS and the origin ? maybe
try removing the cache_range plugin, or
Great... Curious, if you can share.. What was the fix ?
On Fri, Nov 9, 2018 at 8:57 AM Vasanth Mathivanan
wrote:
>
> Yes its problem on application side full content cannot get from there .Now
> its working fine .
>
> Thanks for your support Mile and Jeremy Payne.
>
>
>
patches welcome).
>
> Depending on when you are logging (the hook) for tcpinfo there might be a 1
> to 1 or 1 to many relationship between the tcpinfo logs to access logs.
>
> -Bryan
>
>
>
> On Oct 19, 2017, at 2:45 PM, Jeremy Payne wrote:
>
> Are TCP metrics available to th
Is H2 production ready in ATS 7.5 ? Specifically for the live-linear use case.
Thanks Leif, I'll take that back to my team.
Also.. excuse the typo'd version number.. I was also working on some
RHEL 7.5 issues yesterday.
So some neural cross talk occurring at that time.. :-)
On Mon, Jan 7, 2019 at 9:40 PM Leif Hedstrom wrote:
>
>
>
> > On Jan 7,
re: ATS 7.1.x
Somewhat related.. I notice that upon polling an upstream server, pqsi
does not log the upstream IP.
pqsn DOES log the upstream FQDN.
If I replace pqsi with 'nhi' then I am able to log upstream IP again.
On Mon, Jan 14, 2019 at 1:51 PM Alan Carroll wrote:
>
> One of the issues wi
CONNECT method should work here.. Have you tried that ?
Of course you'll have to explicitly set a proxy at the client end.
On Tue, Feb 19, 2019 at 12:46 PM Eric Chaves wrote:
>
> Hi Folks,
>
> I'm new to traffic-server and I'd like to evaluate it to be used as a
> non-cache forward proxy betwee
question... are you seeing 'document truncated' in the diag log?
On Tue, Feb 19, 2019 at 1:45 PM ezko wrote:
>
> hi,
> with ATS 7.1.6 reverse proxy for live HLS , we are seeing corrupted data on
> the playlist when 300 concurrent clients are playing the same content (same
> profile). we tried sev
t; No , diags looks normal
>
> -----Original Message-
> From: Jeremy Payne [mailto:jp557...@gmail.com]
> Sent: Tuesday, February 19, 2019 10:36 PM
> To: users@trafficserver.apache.org
> Subject: [EXTERNAL] Re: corrupt data with multiple concurrent client
>
> question... are you seein
firm if the ATS knows how to handle the FTP protocol,
> or is it HTTP "aware" only?
>
> Cheers,
>
> Em ter, 19 de fev de 2019 às 17:32, Jeremy Payne
> escreveu:
>>
>> CONNECT method should work here.. Have you tried that ?
>> Of course you'll have to
Is the OS complaining about too many open files?
On Wed, Apr 10, 2019, 8:58 AM ezko wrote:
> Hi,
> we are using ATS 7.1.6 for live HLS/DASH.
> when we start load testing we see that the number of Curr Conn as reported
> by traffic top never goes over 30K (even when there are 60K unique clients)
is ATS under systemd control ?
you may need to set LimitNOFILE in your systemd unit file.
On Wed, Apr 10, 2019 at 9:48 AM ezko wrote:
>
> not that i can tell.
>
> cat /proc/sys/fs/file-max
> 37567508
>
> ulimit -a
> core file size (blocks, -c) unlimited
> data seg size (kbytes
Does netstat show more than 30k established connections?
On Wed, Apr 10, 2019, 10:32 AM ezko wrote:
> it is under systemd
> we tried setting LimitNOFILE in systemd , but it doesn't seem to help
>
>
>
>
>
> --
> Sent from: http://apache-traffic-server.24303.n7.nabble.com/
>
ti-Site Origin configured via proxying as well).
> shn was "origin hostname", pqsn was "proxy hostname", pqsi was "proxy
> resolved IP".
>
> Steve
>
>
> On Wed, Feb 6, 2019 at 11:43 AM Jeremy Payne wrote:
>
>> re: ATS 7.1.x
>>
&g
hange definitively removed/changed the pqsn/pqsi/shn logic that was
> there before when parent.config is in use.
>
> Steve
>
>
> On Tue, Apr 16, 2019 at 8:44 AM Jeremy Payne wrote:
>>
>> Doesnt answer your question as to why the change and documentation
>> discrepa
What about the 'cache promote' plugin ?
On Thu, Jul 25, 2019 at 2:01 PM Sudheer Vinukonda <
sudheervinuko...@yahoo.com> wrote:
> Ah, very interesting. That requires an additional mechanism to maintain
> the state (first or subsequent access) across requests. Yeah, AFAIK, this
> will need a new
Maybe you can also attach strace to a running ATS process. From there you
can see what threads are busy or blocked.
I like to trace on pread64 pwrite64 writev readv and network
On Mon, Sep 9, 2019, 10:54 AM Jason Yang wrote:
> Hi community,
> I am running an ATS cluster with debug workl
Anyone from the community tested BBRv2 in production?
If so, are you seeing any issues, especially with regards to
loss-based congestion control fairness ?
Thanks!
Jeremy
try this..
export http_proxy="http://172.24.0.8:80/";
export https_proxy="http://172.24.0.8:80/";
curl -vvv https://c23-ngxorg-01/assets/sintel/vod/250kbit.m3u8
On Tue, Dec 24, 2019 at 11:41 AM ezko wrote:
>
> Hi,
> we are use ATS 8.0.5 as forward proxy.
> it works fine with HTTP , but fails wi
ATS child/edge is attempting a TLS setup to your parent.
Is the parent configured for TLS communication ?
On Mon, Jan 6, 2020 at 10:38 AM ezko wrote:
>
> hi,
> so trying to configure a simple 2 tier cache with reverse and forward
> caches.
> any ideas why this doesn't work ?
> thanks.
>
> client
Jeremy Payne wrote:
>
> ATS child/edge is attempting a TLS setup to your parent.
> Is the parent configured for TLS communication ?
>
>
> On Mon, Jan 6, 2020 at 10:38 AM ezko wrote:
> >
> > hi,
> > so trying to configure a simple 2 tier cache with reverse and f
Probably the easiest solution is to create another origin fqdn for your
live or vod content.
Then have the corresponding remap use that new fqdn as your origin fqdn
If you cant do that then there are some tricks you can play with the origin
parameters that determine which volume to use. I'm not at
request)
return 0
end
On Sun, Jan 12, 2020 at 12:52 PM Leif Hedstrom wrote:
>
>
>
> On Jan 12, 2020, at 10:47, Jeremy Payne wrote:
>
>
> Probably the easiest solution is to create another origin fqdn for your live
> or vod content.
> Then have the correspondin
You will also have to update your hosting.config to match the volume
where the 'dynamically' set client HOST header should reside.
On Sun, Jan 12, 2020 at 1:16 PM Jeremy Payne wrote:
>
> pulled the below lua from my work notes..
>
> ++
> function send_request()
>
You can see which volume is being used by enabling these debug tags.
cache_hosting|http_seq
On Tue, Jan 14, 2020 at 12:25 PM ezko wrote:
>
> Thanks for the suggestions.
> we are trying with the header rewrite plugin , but can't get it to work
> (even unconditionally):
>
> we enabled pristine: pr
If ATS to origin cert validation is failing, then you may need to
disable server verify(that's if your company security policy allows).
Couple different approaches..
https://docs.trafficserver.apache.org/en/latest/admin-guide/files/sni.yaml.en.html#override-verify-origin-server
https://docs.traf
also.. using curl(without the insecure option), can you poll the
origin server directly without receiving a certificate error ?
On Sun, Mar 1, 2020 at 10:59 AM Jeremy Payne wrote:
>
> If ATS to origin cert validation is failing, then you may need to
> disable server verify(that&
make sure '/opt/ts/etc/ssl/certs/ca.pem' contains the root cert needed
to validate the origin server cert.
also ensure your origin is sending the server cert and appropriate
intermediate cert(s).
On Sun, Mar 1, 2020 at 11:03 AM Jeremy Payne wrote:
>
> also.. using curl(with
gt; certification-path looks good - it uses crt
>
> i converted the crt-file, i received, to pem. maybe that makes a problem.
>
> i will check that.
>
>
> thank you for your help.
>
>
> kind regards
>
>
> Juergen
>
>
>
> Am 01.03.2020 um 19:24 sc
eing used to specify
> Origin failover rules.
>
> Sorry I didn’t mention that at the top.
>
> Thanks
>
> Nick
>
> Sent from my iPhone
>
> On Jun 24, 2020, at 8:59 PM, Jeremy Payne wrote:
>
>
> yeah.. how many parents are listed in parent.config ?
>
objective - expedited(15-20min) draining of segmented video
delivery(live or vod) edge traffic.
(mid/parent traffic draining is fairly simple.. i am only focusing on
edge draining)
so re-visiting draining challenges, specifically with respect to
'sticky clients' and segmented video playback.
anyon
also.. please provide the remap entry in question.
you can sanitize the hostnames.. but keep the scheme references.
i do believe the parent selection must match the origin scheme defined
in the remap.
'502 connection refused' sounds like the parent selection is not
matching and the request is goin
om parent 443 0 no-query tls
>
> But I don't want to use squid.
>
> Thanks again,
> Zack
>
>
> > On Nov 9, 2021, at 11:55 AM, Jeremy Payne wrote:
> >
> > also.. please provide the remap entry in question.
> > you can sanitize the hostnames.. but
re: ATS 9.0.x, 9.1x
I created a list of allowed methods in ip_allow.yaml. With an implicit
DENY:ALL rule.
++
ip_allow:
- apply: in
ip_addrs: 0/0
action: allow
methods:
- GET
- HEAD
- OPTIONS
++
For remaps that need to support POST(or whatever method), I've tried
re: ATS 9.1.2
parent policy = consistent_hash
strategy(nexthop) policy = consistent_hash
num-parent-rings = 2(primary/secondary)
num-nexthop-rings = 2(primary/secondary)
retry-window = 300s
failure-threshold = 10s
parent-connection-timeout = 2s
I notice that the nexthop failure count upon a netwo
re: proxy.config.ssl.client.CA.cert.filename
Our openssl library build does not contain a CA trust/bundle file in
the expected OPENSSLDIR directory.
So I have to set the above ATS parameter so ATS can validate origin
certificates.
The parameter is listed as reloadable in the ATS 9.1.x records.conf
I noticed that parent consistent hashing only works when a unique weight is
defined per parent server; as seen with the below parent.config entry.
dest_domain=. parent="192.168.1.120:80|1.0; 192.168.1.121:180|2.0;
192.168.1.122:80|3.0" round_robin=consistent_hash
If I use either of the below e
In testing TLS connections to my origin complex, I noticed that ATS does
not validate the origin hostname against the server certificate CN/SAN
values.
I then looked at the ATS code that sets the TLS verify options, and noticed
there are no options or routines that validate hostname. So I assume t
Has there been any developments/advancements with respect to purge?
Specifically wildcard purges or generic pattern matching purges.
The cache inspector doesnt scale and the regex_revalidate plugin relies on
user requests combined with a hard set time frame.
So with the regex_reval plugin, I may n
James.. I'll test the genid configuration option. I'll also test writing a
lua script that takes from both the regex_reval and genid plugins.
In this per remap lua script, Ill open a rule file/db The file/db will
contain a list of regex's and a genid. Upon matching, the lua script
will set the cac
BTW, the reason I mention using my own genid in lua, is because I dont see
access to 'proxy.config.http.cache.generation' in the lua plugin code.
Maybe that config over-ride will be added to the Lua plugin in the near
future.
On Mon, Sep 28, 2015 at 7:22 AM, Jeremy Payne wrote:
&
not sure if this answers your question, but i have been thinking about
something similar.
this is how i have it played out in my mind, although i havent looked at
the transaction points to see when host/domain is considered for volume
placement.
1. create a ramdisk(or some memory based filesyetm)
8, 2015, at 12:37 PM, Jeremy Payne wrote:
>
> not sure if this answers your question, but i have been thinking about
> something similar.
>
> this is how i have it played out in my mind, although i havent looked at
> the transaction points to see when host/domain is considered
great.. i am going to test this tonight to see where in the transaction i
can influence the volume placement decision.
On Mon, Sep 28, 2015 at 3:53 PM, Phil Sorber wrote:
> On Mon, Sep 28, 2015 at 2:48 PM Jeremy Payne wrote:
>
>> i am going to test this tonight, but are you sayi
You can also use the lua plugin to set cache-control per remap.
https://github.com/portl4t/ts-lua/wiki/Doc#tsserver_responseheaderheader
On Fri, Oct 2, 2015 at 6:36 AM, tomr wrote:
> Hi All,
>
> I want to use the LHS of the map rule - the pristine host header - as the
> value to a primary speci
Is this happening at the parent or child layer?
At the child, can you relay your remap(if applicable) and parent
configurations?
On Wed, Oct 7, 2015 at 5:49 AM, Randeep wrote:
> Hi,
>
> I have tried configuring a child ats.
>
> In the records.config, I have enabled caching to 1
>
> CONFIG proxy
ote:
> Hi,
>
> This is happening at child end.
>
> I didn't understand the remap part. Remap rules I have configured in the
> parents remap.config.
>
> Regards,
> Randeep
>
> On Wed, Oct 7, 2015 at 7:39 PM, Jeremy Payne wrote:
>
>> Is this happening
Not sure if you are just researching or what.. But this same functionality
is also supported in the lua plugin.
http://trafficserver.readthedocs.org/en/6.0.x/reference/plugins/ts_lua.en.html
ts.client_request.set_url_port
On Wed, Oct 7, 2015 at 12:41 PM, Scott Beardsley
wrote:
> I'd like to mod
Version - ATS 5.3.2
Mode - Reverse Proxy
Anyone have any pointers on configuring ATS to follow HTTP
redirects(301/302) handed by the origin server ?
The goal here is to have ATS follow the redirect on behalf of the client,
instead of sending the redirect back downstream.
This seems doable in Lua(
Tue, Oct 13, 2015 at 4:13 PM, Sudheer Vinukonda
wrote:
> Below are the settings you would need. Both these settings should be
> overridable on master (not sure, about 5.3.x).
>
>
>
>
>
>
> Also, ATS should stream the final (after redirect) response to the end
> user
I see that master(6.1.0) now 'supports' redirection override.
./trafficserver/proxy/http/HttpConfig.cc:
HttpEstablishStaticConfigByte(c.oride.redirection_enabled,
"proxy.config.http.redirection_enabled");
./trafficserver/proxy/http/HttpConfig.cc:
params->oride.redirection_enabled =
INT_TO_BOOL(m_m
, Sudheer Vinukonda
wrote:
> Which hook are you using to override?
>
> The config is applied before READ_REQUEST_HDR hook, so, you will need to
> override it in TXN_START hook.
>
> Thanks,
>
> Sudheer
>
>
>
> On Wednesday, November 11, 2015 10:12 AM, Jeremy Payne
t;
> > Thanks,
> >
> > Sudheer
> >
> >
> >
> > On Wednesday, November 11, 2015 10:12 AM, Jeremy Payne <
> jp557...@gmail.com> wrote:
> >
> >
> >
> >
> > I see that master(6.1.0) now 'supports' redirection override.
&
Jira opened for this work.
https://issues.apache.org/jira/browse/TS-4017
"Objective here is to allow proxy.config.http.redirection_enabled override
to be set in all transaction hooks up to and including 'READ_RESPONSE_HDR'"
On Wed, Nov 11, 2015 at 5:49 PM, Jeremy Payne wr
If you are looking to change User-Agent based on URL and User-Agent then
you may want to look at either of the below two plugins.
https://docs.trafficserver.apache.org/ja/latest/reference/plugins/ts_lua.en.html
https://docs.trafficserver.apache.org/ja/latest/reference/plugins/header_rewrite.en.ht
So you see the SYN from the client hit the ATS server?
But the ATS machine doesnt respond with the SYN-ACK ?
As a start and on the ATS machine can you try listening on all interfaces?
tcpdump -ni any port 80 or port 8080
On Fri, Nov 20, 2015 at 2:43 AM, Wayne Zhang
wrote:
> I want to correc
This is doable with the lua plugin.
On Thu, Dec 3, 2015 at 7:51 AM, Steve Malenfant
wrote:
> Is there way we can modify the path of a filename requested to an origin?
> In this example, I want to remove the "path" portion and only send the
> "filename" to the origin.
>
>
> Looks like regex_map w
I noticed that the 'proxy.config.http.background_fill_active_timeout'
configuration option
is not honored upon a pre-mature client close. Debug shows
that the option value is set, but in practice the origin connection
remains open even after the set timeout value.
Maybe I need to set another op
1 - 100 of 136 matches
Mail list logo
