openwhisk-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Kumar Subramanian <kumarsubr...@vmware.com>
Subject Re: Error when deploying Whisk-Controller in dcos
Date Mon, 05 Mar 2018 22:58:05 GMT
Hi Carlos/Tyson/Chetan,
Any Suggestions?

Note: I just tried to increase the timeout on the health checks …no luck.

Thanks,
Kumar. 

On 3/5/18, 12:28 PM, "Kumar Subramanian" <kumarsubrama@vmware.com> wrote:

    This is the output log
    [2018-03-05T20:24:10.016Z] [INFO] Initializing Kamon...
    [INFO] [03/05/2018 20:24:10.301] [main] [StatsDExtension(akka://kamon)] Starting the Kamon(StatsD)
extension
    [2018-03-05T20:24:10.352Z] [INFO] Slf4jLogger started
    [2018-03-05T20:24:10.706Z] [INFO] [??] [Config] environment set value for db.whisk.actions
    [2018-03-05T20:24:10.707Z] [INFO] [??] [Config] environment set value for db.protocol
    [2018-03-05T20:24:10.708Z] [INFO] [??] [Config] environment set value for limits.actions.sequence.maxLength
    [2018-03-05T20:24:10.708Z] [INFO] [??] [Config] environment set value for limits.triggers.fires.perMinute
    [2018-03-05T20:24:10.708Z] [INFO] [??] [Config] environment set value for akka.cluster.seed.nodes
    [2018-03-05T20:24:10.709Z] [INFO] [??] [Config] environment set value for limits.actions.invokes.concurrent
    [2018-03-05T20:24:10.709Z] [INFO] [??] [Config] environment set value for controller.instances
    [2018-03-05T20:24:10.710Z] [INFO] [??] [Config] environment set value for controller.localBookkeeping
    [2018-03-05T20:24:10.710Z] [INFO] [??] [Config] environment set value for whisk.version.date
    [2018-03-05T20:24:10.710Z] [INFO] [??] [Config] environment set value for db.port
    [2018-03-05T20:24:10.711Z] [INFO] [??] [Config] environment set value for whisk.version.buildno
    [2018-03-05T20:24:10.711Z] [INFO] [??] [Config] environment set value for db.whisk.activations
    [2018-03-05T20:24:10.711Z] [INFO] [??] [Config] environment set value for db.username
    [2018-03-05T20:24:10.712Z] [INFO] [??] [Config] environment set value for limits.actions.invokes.perMinute
    [2018-03-05T20:24:10.712Z] [INFO] [??] [Config] environment set value for db.whisk.auths
    [2018-03-05T20:24:10.712Z] [INFO] [??] [Config] environment set value for limits.actions.invokes.concurrentInSystem
    [2018-03-05T20:24:10.712Z] [INFO] [??] [Config] environment set value for runtimes.manifest
    [2018-03-05T20:24:10.713Z] [INFO] [??] [Config] environment set value for kafka.hosts
    [2018-03-05T20:24:10.713Z] [INFO] [??] [Config] environment set value for db.host
    [2018-03-05T20:24:10.713Z] [INFO] [??] [Config] environment set value for port
    [2018-03-05T20:24:10.714Z] [INFO] [??] [Config] environment set value for db.password
    [2018-03-05T20:24:10.714Z] [INFO] [??] [Config] environment set value for db.provider
    Received killTask for task whisk-controller.28d25d91-20b3-11e8-8754-3afdc003616b
    [INFO] [03/05/2018 20:25:38.974] [kamon-shutdown-hook-1] [CoordinatedShutdown(akka://kamon)]
Starting coordinated shutdown from JVM shutdown hook
    [2018-03-05T20:25:38.975Z] [INFO] Starting coordinated shutdown from JVM shutdown hook
    [2018-03-05T20:25:38.982Z] [INFO] [??] [Controller] Shutting down Kamon with coordinated
shutdown
    
    ERROR_LOG
    I0305 20:24:09.220711  1337 exec.cpp:162] Version: 1.2.3
    I0305 20:24:09.227144  1338 exec.cpp:237] Executor registered on agent 995020e0-5129-44a3-8cf4-65900838b3af-S7
    W0305 20:24:09.227144  1341 logging.cpp:91] RAW: Received signal SIGTERM from process
10243 of user 0; exiting
    
    On 3/5/18, 12:24 PM, "Kumar Subramanian" <kumarsubrama@vmware.com> wrote:
    
        I get the following error while in the deploying state (then kills it automatically
and re-installs and goes on…)
        Task was killed since health check failed. Reason: ConnectionAttemptFailedException:
Connection attempt to <whisk_controller_ip>>:8888 failed
        
        
        
        On 3/5/18, 12:21 PM, "Kumar Subramanian" <kumarsubrama@vmware.com> wrote:
        
            Gave the value as mykafka.marathon.mesos:9092…it seems to be going forward now
with the deployment…hope it succeeds
            
            On 3/5/18, 12:13 PM, "Kumar Subramanian" <kumarsubrama@vmware.com> wrote:
            
                Hi Chetan,
                I resolved all the settings (not sure about some those values set)….now
I’m getting the following error
                I0305 20:07:00.341622 21216 exec.cpp:162] Version: 1.2.3
                I0305 20:07:00.347102 21224 exec.cpp:237] Executor registered on agent 995020e0-5129-44a3-8cf4-65900838b3af-S4
                Exception in thread "main" org.apache.kafka.common.KafkaException: Failed
create new KafkaAdminClient
                	at org.apache.kafka.clients.admin.KafkaAdminClient.createInternal(KafkaAdminClient.java:322)
                	at org.apache.kafka.clients.admin.AdminClient.create(AdminClient.java:50)
                	at whisk.connector.kafka.KafkaMessagingProvider$.ensureTopic(KafkaMessagingProvider.scala:70)
                	at whisk.core.controller.Controller$.main(Controller.scala:217)
                	at whisk.core.controller.Controller.main(Controller.scala)
                Caused by: org.apache.kafka.common.config.ConfigException: No resolvable bootstrap
urls given in bootstrap.servers
                	at org.apache.kafka.clients.ClientUtils.parseAndValidateAddresses(ClientUtils.java:64)
                	at org.apache.kafka.clients.admin.KafkaAdminClient.<init>(KafkaAdminClient.java:345)
                	at org.apache.kafka.clients.admin.KafkaAdminClient.createInternal(KafkaAdminClient.java:315)
                	... 4 more
                <<WHISK_CONTROLLER ENVIRONMENT SETTINGS>>
                The config environment values are:
                "KAFKA_HOST": "broker-0.kafka.mesos"
                "KAFKA_HOSTS": "mykafka.docker:9092"
                
                
                Here is the Kakfka service in my dcos services
                <<KAFKA_SERVICE>>	
                ID: mykafka.654fa8cd-1e56-11e8-8754-3afdc003616b
                Name: mykafka
                Address: <<internal_ip>>	
                Status: Running
                
                On 3/5/18, 12:01 PM, "Kumar Subramanian" <kumarsubrama@vmware.com> wrote:
                
                    Also what is DB_WHISK_ACTIVATIONS=local_activations, what does it mean
by “local” activations? This setting is also needed, what should the value be in my dcos
env? Should it still be local_activations?
                    
                    On 3/5/18, 11:48 AM, "Kumar Subramanian" <kumarsubrama@vmware.com>
wrote:
                    
                        Thanks Chetan, I added that now I’m getting the 
                        
                        [2018-03-05T19:43:45.025Z] [ERROR] [??] [Config] required property
kafka.hosts still not set
                        
                        what is kafka.hosts is that the kafka host name? (as in mykafka…that;s
the kafka name I gave when I installed kakfa). Should it be just the name or is FQDN ?
                         
                        
                        On 3/5/18, 11:32 AM, "Chetan Mehrotra" <chetan.mehrotra@gmail.com>
wrote:
                        
                            > required property controller.instances still not set
                            
                            Looks like some configs are missing. You would need to this or
few
                            more props. The configs are generally managed via Ansible for
default
                            setup. For dcos you may need to configure them explicitly. You
can see
                            various configs and there values as an example at [1]
                            (controller.instances becomes CONTROLLER_INSTANCES). They would
need
                            to be tweaked as per your setup though
                            
                            Chetan Mehrotra
                            [1] https://urldefense.proofpoint.com/v2/url?u=https-3A__github.com_apache_incubator-2Dopenwhisk-2Ddevtools_blob_master_docker-2Dcompose_docker-2Dwhisk-2Dcontroller.env&d=DwIFaQ&c=uilaK90D4TOVoH58JNXRgQ&r=F5C8fYlpBJ270qrdwLq2iRQrPd1CLap8zItxk8laWpo&m=K8Rzl5BrVqOWFx5b1fYG8EjdY-6JVyi-x_eMD0thaKY&s=VfbixlUG4tFbgBNxiMr-KCNJOakmlGuNcnXLlbQMrVY&e=
                            
                            
                            On Tue, Mar 6, 2018 at 12:09 AM, Kumar Subramanian
                            <kumarsubrama@vmware.com> wrote:
                            > Hi,
                            > I was able to successfully do the following
                            > 1) Build the Controller image
                            > 2) Push the image
                            >
                            > However when I installed the controller package it gives
me the following error in the output; then it shuts down and retries the installation (goes
on…)
                            >
                            > Registered docker executor on 10.0.6.13
                            > Starting task whisk-controller.a46ae612-20a2-11e8-8754-3afdc003616b
                            > [2018-03-05T18:25:55.853Z] [INFO] Initializing Kamon...
                            > [INFO] [03/05/2018 18:25:56.151] [main] [StatsDExtension(akka://kamon)]
Starting the Kamon(StatsD) extension
                            > [2018-03-05T18:25:56.193Z] [INFO] Slf4jLogger started
                            > [2018-03-05T18:25:56.552Z] [INFO] [??] [Config] environment
set value for db.whisk.actions
                            > [2018-03-05T18:25:56.554Z] [INFO] [??] [Config] environment
set value for db.protocol
                            > [2018-03-05T18:25:56.554Z] [INFO] [??] [Config] environment
set value for limits.triggers.fires.perMinute
                            > [2018-03-05T18:25:56.554Z] [INFO] [??] [Config] environment
set value for limits.actions.invokes.concurrent
                            > [2018-03-05T18:25:56.555Z] [INFO] [??] [Config] environment
set value for whisk.version.date
                            > [2018-03-05T18:25:56.555Z] [INFO] [??] [Config] environment
set value for db.port
                            > [2018-03-05T18:25:56.555Z] [INFO] [??] [Config] environment
set value for whisk.version.buildno
                            > [2018-03-05T18:25:56.556Z] [INFO] [??] [Config] environment
set value for db.username
                            > [2018-03-05T18:25:56.556Z] [INFO] [??] [Config] environment
set value for limits.actions.invokes.perMinute
                            > [2018-03-05T18:25:56.556Z] [INFO] [??] [Config] environment
set value for db.whisk.auths
                            > [2018-03-05T18:25:56.556Z] [INFO] [??] [Config] environment
set value for limits.actions.invokes.concurrentInSystem
                            > [2018-03-05T18:25:56.557Z] [INFO] [??] [Config] environment
set value for runtimes.manifest
                            > [2018-03-05T18:25:56.557Z] [INFO] [??] [Config] environment
set value for db.host
                            > [2018-03-05T18:25:56.558Z] [INFO] [??] [Config] environment
set value for port
                            > [2018-03-05T18:25:56.558Z] [INFO] [??] [Config] environment
set value for db.password
                            > [2018-03-05T18:25:56.558Z] [INFO] [??] [Config] environment
set value for db.provider
                            > [2018-03-05T18:25:56.561Z] [ERROR] [??] [Config] required
property controller.instances still not set
                            > [2018-03-05T18:25:56.561Z] [ERROR] [??] [Controller] Bad
configuration, cannot start.
                            >
                            > Any suggestions?
                            >
                            >
                            > On 3/2/18, 4:05 PM, "Kumar Subramanian" <kumarsubrama@vmware.com>
wrote:
                            >
                            >     This is the error I get when I did docker build (for
Controller)
                            >
                            >     Step 1/7 : FROM scala
                            >     repository scala not found: does not exist or no pull
access
                            >
                            >
                            >     Any Suggestions?
                            >
                            >     On 3/2/18, 3:34 PM, "Carlos Santana" <csantana23@gmail.com>
wrote:
                            >
                            >         No that it’s still in PR
                            >
                            >         Just pull the changes locally and build
                            >
                            >         - Carlos Santana
                            >         @csantanapr
                            >
                            >         > On Mar 2, 2018, at 6:20 PM, Kumar Subramanian
<kumarsubrama@vmware.com> wrote:
                            >         >
                            >         > Is that change at https://urldefense.proofpoint.com/v2/url?u=https-3A__github.com_chetanmeh_incubator-2Dopenwhisk_blob_fa302249f4f9b4e6b3084956f18bda987674f46f_core_controller_Dockerfile&d=DwIFaQ&c=uilaK90D4TOVoH58JNXRgQ&r=F5C8fYlpBJ270qrdwLq2iRQrPd1CLap8zItxk8laWpo&m=nLubLAFijdQ4pOPqIydDI_wguMgbdmdmoMXcP7g-m8k&s=0_zv4jTDip5Uk9oBB5-6Ka_Iug3KYWIhy7qzSDryqM0&e=
not merged? I don’t see change in master https://urldefense.proofpoint.com/v2/url?u=https-3A__github.com_apache_incubator-2Dopenwhisk_blob_master_core_controller_Dockerfile&d=DwIFaQ&c=uilaK90D4TOVoH58JNXRgQ&r=F5C8fYlpBJ270qrdwLq2iRQrPd1CLap8zItxk8laWpo&m=nLubLAFijdQ4pOPqIydDI_wguMgbdmdmoMXcP7g-m8k&s=TdB-IjihM3-0dqBF029dSMkoWbZHBAUXXqeDjfQMlhg&e=
                            >         >
                            >         > Thanks,
                            >         > Kumar.
                            >         >
                            >         > On 3/2/18, 2:59 PM, "Kumar Subramanian" <kumarsubrama@vmware.com>
wrote:
                            >         >
                            >         >    Ok, I will try to build the controller image
and see. Will keep you posted.
                            >         >
                            >         >    On 3/2/18, 2:39 PM, "Tyson Norris" <tnorris@adobe.com.INVALID>
wrote:
                            >         >
                            >         >        Thanks Carlos - I think you’re right.
                            >         >
                            >         >
                            >         >
                            >         >        Kumar you can either build the controller
image with that PR, or else you should be able to manually set the docker cmd, e.g. /bin/sh
-c \"exec /init.sh 0 >> /dev/stdout\” on the dcos service for controller;
                            >         >
                            >         >
                            >         >
                            >         >        I think you will have similar issue with
invoker, mostly because this universe is far out of date from current openwhisk images.
                            >         >
                            >         >
                            >         >
                            >         >        For invoker can you use the docker cmd
as /bin/sh -c \"exec /init.sh --name $LIBPROCESS_IP >> /dev/stdout\”
                            >         >
                            >         >
                            >         >
                            >         >        Additionally, the env vars (both invoker
and controller) have changed substantially, so I would expect a few hiccups there as well.
                            >         >
                            >         >
                            >         >
                            >         >        We are working on getting updates to
the universe so that our internal deployment details are not included, and it will actually
work with recent openwhisk images (and stay working) but haven’t gotten everything set just
yet.
                            >         >
                            >         >
                            >         >
                            >         >        Hope that helps
                            >         >
                            >         >        Tyson
                            >         >
                            >         >
                            >         >
                            >         >>> On Mar 2, 2018, at 2:21 PM, Carlos Santana
<csantana23@gmail.com> wrote:
                            >         >>
                            >         >>
                            >         >
                            >         >> Maybe for the init.sh this PR is related
                            >         >
                            >         >> https://urldefense.proofpoint.com/v2/url?u=https-3A__na01.safelinks.protection.outlook.com_-3Furl-3Dhttps-253A-252F-252Fgithub.com-252Fapache-252Fincubator-2Dopenwhisk-252Fpull-252F3374-252Ffiles-2523diff-2D8f445fbdf6253dd176975ff6c629def4R18-26data-3D02-257C01-257Ctnorris-2540adobe.com-257C605cc50c7a484ccb833708d5808bfe64-257Cfa7b1b5a7b34438794aed2c178decee1-257C0-257C0-257C636556261107863657-26sdata-3Dtk8Des10hubLs7FNSgzDlsk1ibxDTIqSlXti-252FcAUyz0-253D-26reserved-3D0&d=DwIGaQ&c=uilaK90D4TOVoH58JNXRgQ&r=F5C8fYlpBJ270qrdwLq2iRQrPd1CLap8zItxk8laWpo&m=LUthdew4Dt10vSAZSYRBbREqgwWk2PUWc4KDBJtt0uU&s=3UqWljTQjItMnzWhPrsfD1AF2IX6abtc9dYRfrxb2_M&e=
                            >         >
                            >         >>
                            >         >
                            >         >>
                            >         >
                            >         >> On Fri, Mar 2, 2018 at 5:09 PM Tyson Norris
<tnorris@adobe.com.invalid>
                            >         >
                            >         >> wrote:
                            >         >
                            >         >>
                            >         >
                            >         >>> Check your marathon/dcos service config
to verify what image is used, and
                            >         >
                            >         >>> that you have the latest image pulled?
                            >         >
                            >         >>>
                            >         >
                            >         >>> The default should be openwhisk/controller
- but I see that universe
                            >         >
                            >         >>> package marathon config is not set to
force pull, so if you are using that
                            >         >
                            >         >>> image, make sure you have pulled the
latest manually (or change the config
                            >         >
                            >         >>> to force pull in dcos/marathon ui).
                            >         >
                            >         >>>
                            >         >
                            >         >>> Tyson
                            >         >
                            >         >>>
                            >         >
                            >         >>>> On Mar 2, 2018, at 12:45 PM, Kumar
Subramanian <kumarsubrama@vmware.com>
                            >         >
                            >         >>> wrote:
                            >         >
                            >         >>>>
                            >         >
                            >         >>>> Hi,
                            >         >
                            >         >>>> I have installed the following in
DCOS successfully:
                            >         >
                            >         >>>> 1. Apigateway
                            >         >
                            >         >>>> 2. Exhibitor-dcos
                            >         >
                            >         >>>> 3. Kafka (name given is mykafka
at the time of installation)
                            >         >
                            >         >>>> 4. Whisk-couchdb
                            >         >
                            >         >>>> 5. Consul
                            >         >
                            >         >>>> 6. Registrator
                            >         >
                            >         >>>>
                            >         >
                            >         >>>> Eror when deploying Whisk-Controller
in DCOS:
                            >         >
                            >         >>>> When I tried to deploy whisk-controller
with default settings, then the
                            >         >
                            >         >>> service fails to deploy (it just kills
and redploys the service
                            >         >
                            >         >>> continuously on its own when deploying)
                            >         >
                            >         >>>>
                            >         >
                            >         >>>> Here is the content in the Error
and Output
                            >         >
                            >         >>>>
                            >         >
                            >         >>>> STDERR:
                            >         >
                            >         >>>> (AT BEGINNING OF FILE)
                            >         >
                            >         >>>> I0302 20:38:35.176177 19822 exec.cpp:162]
Version: 1.2.3
                            >         >
                            >         >>>> I0302 20:38:35.180703 19824 exec.cpp:237]
Executor registered on agent
                            >         >
                            >         >>> 995020e0-5129-44a3-8cf4-65900838b3af-S6
                            >         >
                            >         >>>> docker: Error response from daemon:
Container command 'init.sh' not
                            >         >
                            >         >>> found or does not exist..
                            >         >
                            >         >>>>
                            >         >
                            >         >>>> OUTPUT:
                            >         >
                            >         >>>> (AT BEGINNING OF FILE)
                            >         >
                            >         >>>> Registered docker executor on 10.0.6.55
                            >         >
                            >         >>>> Starting task whisk-controller.adb62c44-1e59-11e8-8754-3afdc003616b
                            >         >
                            >         >>>>
                            >         >
                            >         >>>> Can you please provide your valuable
inputs on how to get
                            >         >
                            >         >>> whisk-controller deployed in dcos?
                            >         >
                            >         >>>>
                            >         >
                            >         >>>> Thanks,
                            >         >
                            >         >>>> Kumar.
                            >         >
                            >         >>>>
                            >         >
                            >         >>>
                            >         >
                            >         >>>
                            >         >
                            >         >
                            >         >
                            >         >
                            >         >
                            >         >
                            >         >
                            >
                            >
                            >
                            >
                            
                        
                        
                    
                    
                
                
            
            
        
        
    
    

Mime
View raw message