I have a client application that use the REST API, and by default is
specifies a clientId parameter on the URL. When my client terminates
abruptly without a request having the "&action=unsubscribe" parameter, then
a consumer can be orphaned on the server/broker. Since an orphaned consumer
can caus
The following is a sample SAS program that puts a message on a queue, then
gets it off and writes it to a flat file:
/* Assign a fileref to the queue. */
filename myq ActiveMQ user=admin pass=admin
URL='http://localhost:8161/api/message/mytestq?type=queue';
/* Put a message on a queue named "
I also tried:
-
Stephen Vincent
--
View this message in context:
http://activemq.2283324.n4.nabble.com/REST-API-client-ecountering-Status-401-Unauthorized-tp4666184p4666187.html
Sent from the ActiveMQ - User mailing list archive at Nabble.com.
In http://activemq.apache.org/security.html, in the "Authorization Example"
section it says: "The following example shows these 2 plugins in
operation." The word "example" in that is a link to:
http://svn.apache.org/repos/asf/activemq/trunk/activemq-core/src/test/resources/org/apache/active
I have also tried the userid in the form of \ but still get
unauthorized.
-
Stephen Vincent
--
View this message in context:
http://activemq.2283324.n4.nabble.com/REST-API-client-ecountering-Status-401-Unauthorized-tp4666184p4666185.html
Sent from the ActiveMQ - User mailing list archive at
I'm attempting to run a secure broker using release 5.8. on Windows. I was
able to make it work when the broker was on a UNIX machine. But I have
something incorrect when trying to configure my PC. My REST API client
encounters:
Status 401 -- Unauthorized.
Unauthorized: Check userna
I feel sure that 5.9 invocation failure I mentioned in my last post was
caused by an incomplete extraction of the tar file. I was testing on an
UNIX box running HP-UX 11.31 and its tar command is old. It did not
accommodate the long path names. So, I downloaded the 04Apr 5.9 snapshot
and used gt
I attempted to test it with the 3/27 5.9 snapshot on a UNIX machine running
HP-UX 11.31, but it failed to start due to:
Loading message broker from: xbean:activemq.xml
INFO | Refreshing org.apache.activemq.xbean.XBeanBrokerFactory$1@89cc5e:
startup date [Mon Apr 01 12:57:29 EDT 2013]; root of con
Yes, I am still seeing problems. I have tested both 5.7 and 5.8. Sometimes,
messages are de-queued but don't make it to my RESTful client. I don't know
what happens to the messages when this occurs. I assume the consumer
created in the broker's JVM may get them, but my RESTful client's GET
requ
Q1: When an AMQP client puts a text message on a destination, does it get
converted into a JMS TextMessage?
Q2: Can an AMQP client get a JMS TextMessage off of an ActiveMQ destination?
I ask those questions because I wonder if the following steps are possible:
1. AMQP client puts a text messag
Yes, I am using the default kahaDB as the broker store.
My tests seem to run fine if I access a broker that is running on my PC, and
I specify the prioritizedMessages="true" policy using the queue name instead
of the ">" wildcard. However, they frequently encounter trouble when the
broker is runn
That last post didn't give you an idea of how many messages it takes before I
see trouble. I have seen it fail after just a few iterations. So, I have
seen it fail with under 200 messages.
-
Stephen Vincent
--
View this message in context:
http://activemq.2283324.n4.nabble.com/Does-the-d
No, just 13 messages, but I run the test over and over again. Most of the
time it is successful, but sometimes fails. It fails about 10 to 20% of the
time when the broker is running on a 64-bit UNIX machine running HP-UX
11.31.
The test first POSTs two messages with the default priority. Then P
Running release 5.7.0.
I have some REST API client tests that run fine most of the time. However,
we have a few tests that fail roughly 10% of the time when run against a
server running on UNIX. The following are excerpts from the log. I see a
bunch of these:
---
2013-02-04 13:02:50,85
If I abruptly kill my REST API client, I am sometimes left with an orphaned
consumer over on the server. Aside from bouncing the server/broker, how do
I kill those left-over consumers? The web admin interface doesn't seem to
provide a way.
-
Stephen Vincent
--
View this message in context:
Well, my tester has encountered more prioritized message queue test failures
even though she removed those prefetch=0 policies from the activemq.xml
config file, and the failures are resolved if she uses a queue name when
specifying the prioritizedMessages="true" policy, instead of attempting to
u
In summary, my test involves a queue where prioritizedMessages="true". I put
some messages with varying priorities on the queue, and then GET them off to
see if they are retrieved in order of highest priority. My test as a loop
doing that many times. When prefetch is zero, it doesn't take many
i
More testing indicates that the ">" by itself seems to be ok. It appears
that my test fails when the following is in the "activemq.xml" file:
If I remove those two lines, and keep:
then my tests run fine.
-
Stephen Vincent
--
View this message in context:
http://activemq
Version 5.7.0
If it matters, I am specifying the destination policy in the
"conf/activemq.xml" file that is being read when the server is started.
When server is running on Windows, I don't see the problem. However, when I
run the server on Solaris, I can reproduce the problem.
-
Stephe
Regarding the ">" destination
wildcard(http://activemq.apache.org/wildcards.html), I see an example in
http://activemq.apache.org/how-can-i-support-priority-queues.html where is
shows the use of just ">" when setting a destination policy for all queues.
Specifically, the example shows:
Some
I ran the test against a 21Dec2012 5.8 snapshot and the problem was not
resolved. I will try a more recent snapshot when I gets some more time.
-
Stephen Vincent
--
View this message in context:
http://activemq.2283324.n4.nabble.com/RESTful-POSTing-of-messages-containing-a-correlation-ID-t
I was able to update my C-based client to add/update the URI in HTTP
start-line(first header). When the user of my client app wants to set the
JMSCorrelationID header field of a message, I add a "&JMSCorrelationID="
parameter to the URI in the HTTP POST request's start-line(first header).
That en
I am re-wording the previous post since it may not be clear.
Within the scope of a REST client session, the first "selector:" header I
send with a GET request is honored. However, selector headers in subsequent
GET requests within the session are ignored in favor of the first selector
header that
I looked deeper into "MessageServlet.java" and see the getReadTimeout changes
readTimeout values that are less than zero or greater than
maximumReadTimeout to maximumReadTimeout. That explains why I was seeing a
20 second timeout. In my client application, I plan to set a default
time-out value o
I see defaultReadTimeout=-1 in "MessageServlet.java", but when my client does
not specify a readTimeout= parameter on the URL, and there are no messages
on the queue, then my HTTP client actually experiences a timeout of 20
seconds. Is that a bug? Am I looking at the wrong code?
That 20 second t
I found that my client was sending way more than it was supposed to be
sending in the POST request. I fixed my code and now can successfully run
my client on z/OS.
-
Stephen Vincent
--
View this message in context:
http://activemq.2283324.n4.nabble.com/HttpException-413-FULL-head-null-at-H
When using the RESTful HTTP interface and I make a GET request, the message
is sent back with a linefeed character appended at the end. My broker is
running on Windows. Will this always be the case?
-
Stephen Vincent
--
View this message in context:
http://activemq.2283324.n4.nabble.com/R
Within a client session, the first "selector:" header I send with a GET
request is honored, but I am unable to change that selector on subsequent
GET requests within the same session. Am I out-of-luck, or is there a way to
have subsequent GET requests on the same session(having different selector
v
2012-10-23 18:53:34,270 | DEBUG | HttpParser{s=-14,l=0,c=-3} |
org.eclipse.jetty.http.HttpParser | qtp21584824-33
HttpException(413,FULL head,null)
at org.eclipse.jetty.http.HttpParser.fill(HttpParser.java:1030)
at org.eclipse.jetty.http.HttpParser.parseNext(HttpParser.java:280)
If my HTTP client were to make a POST request containing a header such as
"JMSCorrelationID: mycorelID", it would be great if the MessageServlet would
notice it and call message.setJMSCorrelationID. Is that something that could
be considered? I would love to be able to make many POST requests withi
I realized that not only did my test pgm have an extra post increment of a
loop variable, but it was setting a string property named correlationId as
well as setting the correlation ID header field. Specifically:
message.setStringProperty( "correlationId", correlId );
message.setJMSCorrelati
I believe the JMSCorrelationID is a message header field, not a message
property. While I see the following code excerpt from MessageServlet.java
spins through the message properties and adds HTTP response headers, I don't
believe that includes the JMSCorrelationID.
-
protected void setResp
I downloaded the apache-activemq-5.7-20120922.031357-5-bin.zip file,
extracted it, invoked bin\activmq, but I still don't see that it is fixed.
Maybe I am doing something wrong. Did you try the java test pgm I attacted
to the jira? Does it work for you? I still do not see a correlation header
i
I created "https://issues.apache.org/jira/browse/AMQ-4072";. It would be
great if a fix can make it in 5.7. If there is a circumvention, please let
me know.
Thanks in advance.
-
Stephen Vincent
--
View this message in context:
http://activemq.2283324.n4.nabble.com/No-correlation-ID-in-HT
I modified the test program I included in my last post, so that it prints the
headers. I believe this test illustrates that the correlation ID is not
coming back in an HTTP header. When I run the test below, here is an
example of what it prints:
Set-Cookie: JSESSIONID=b27ojdk
Maybe I am missing something, but the test mentioned seems to get and compare
the body of the message with the correlation ID string that was also sent as
the body/content of the message. In other words, I believe that test
receives the correlation ID by way of the body because that is the content
I have also tried it with a snap-shot of 5.7, but I still do not see a
correlation ID header. Is there some configuration I need?
Thanks,
sv
-
Stephen Vincent
--
View this message in context:
http://activemq.2283324.n4.nabble.com/No-correlation-ID-in-HTTP-header-when-using-RESTful-interfa
Thanks Gary. I do appreciate the help.
I assume it is ok to specify a charset= sub-option following "text/xml".
For example, I have a test case where I specify "text/xml;charset=UTF-8",
and it seems to work. Is the list of supported encodings restricted?
The only way I have been able
I can use the RESTful HTTP protocol to POST and GET messages as long as I
POST using either "text/xml" or "application/x-www-form-urlencoded" as the
Content-Type. If I try to POST via "text/plain", messages are put on the
queue, but have no content. When I attempt to GET those messages off of the
I don't think it was the encoding issue. I now suspect it encounters trouble
in a low-memory situation. I have a test case I run on my PC that POSTs a
number of messages and then GETs them back off of the queue. If I keep the
number of messages under 650, I can run it over and over again without
When I make a HTTP POST request, the "messageID" header is the second header
in the response. When I make a HTTP GET request, the "ID" header is the
fourth header in the response. Will they always have that name and be in
that order? Or should I program my HTTP client to check every header when
I believe that problem was caused by me specifying the wrong character
encoding.
-
Stephen Vincent
--
View this message in context:
http://activemq.2283324.n4.nabble.com/ActiveMQ-5-6-IllegalStateException-MessageServlet-java-368-tp4651797p4652135.html
Sent from the ActiveMQ - User mailing lis
I attempt to put a text message on a queue using HTTP. The admin shows the
enqueued *and* dequeued counters both incremented by one. The broker log
excerpt follows. I must be doing something it doesn't like. Any assistance
will be apreciated. Thanks.
2012-05-23 12:09:56,040 | DEBUG | call se
I can successfully put and get a message using my client app by way of HTTP
if I change the Content-Type from text/plain to
"application/x-www-form-urlencoded" and precede my body with
"destination=request&type=queue&body=".
The default demo web app doesn't seem to like the way I tried to send
tex
It appears that the message did not get my content. The following is a debug
excerpt from the broker log. Please let me know if it reveals a mistake on
my end. Thanks in advance.
2012-05-08 12:29:10,308 | DEBUG | created
SCEP@1efb0be{l(/127.0.0.1:53827)<->r(/127.0.0.1:8161),d=false,open=true,is
I just tried it using ActiveMQ 5.6.0, but I still don't see any message body.
When I browse the message via
"ocalhost:8161/admin/message.jsp?id=ID:d22275-4570-1336415971483-3:1:1:1:1&JMSDestination=request",
the "Type" is blank even though my POST does specify "Content-Type:
text/plain; charset=wi
My JMS java client creates one consumer with a message selector and works
fine when receiving messages from WebSphere MQ. But seldom works when
receiving from a queue on ActiveMQ 5.3.1 and 5.4.1. Most of the time I
don't get any messages. I can consistently get it to work *IF* I am
stepping th
I posted to the ActiveMQ - Dev forum, but the post has not yet been sent to
the maillist. I am having trouble subscribing. Since the addressee email
address is too long, I did what the email instructed and sent email to
"dev-allow-requ...@activemq.apache.org" with the subject of
"dev-allow-uc.12
48 matches
Mail list logo
