Hi,
For slider version 0.92, on the download side, I could only find sources. I
could not find full assemblies (like version 0.80 etc.).
Thoughts?
Thanks,
Thanks Gaur for prompt reply !
Manoj
On Mon, Apr 10, 2017 at 5:32 PM, Gour Saha wrote:
> https://repository.apache.org/content/groups/public/org/
> apache/slider/slide
> r-assembly/0.92.0-incubating/
>
>
> On 4/10/17, 3:00 PM, "Manoj Samel" wrote:
>
> >H
t;
> On 4/10/17, 3:00 PM, "Manoj Samel" wrote:
>
> >Hi,
> >
> >For slider version 0.92, on the download side, I could only find sources.
> >I
> >could not find full assemblies (like version 0.80 etc.).
> >
> >Thoughts?
> >
> >Thanks,
>
>
Hi
Running slider 0.92 on CDH 5.5.1 (which is Hadoop 2.6), with Kerberos
I am deploying a application with multiple components. The components start
but fail to heart beat to slider AM. The slider AM log shows NPE at
container heartbeat URLs as below.
I have attached the complete slider AM log
eve I've come across this error before. The problem was that the
> metainfo.xml file in the application package I was trying to deploy was
> malformed / missing required tags.
>
>
> On 12 April 2017 at 02:09, Manoj Samel wrote:
>
> > Hi
> >
> > Runnin
Slider 0.92 on secured hadoop 2.6 cluster
Have a app with component "tenant1" running. Add another component
"tenant2" and do a upgrade. After that, the original component keeps
showing "awaiting heartbeat..." in the output of "slider list
--containers" ... The component and AM log does not seem
not be used in any role names; then slider 0.92 should give error
when creating cluster or accepting configs during any other operations etc.
saying invalid role name etc. etc.
Thanks in advance,
On Tue, Apr 11, 2017 at 6:09 PM, Manoj Samel
wrote:
> Hi
>
> Running slider 0.92 on CDH 5.5.1
e.
>
> Yeah, we could think about possible ways of solving this problem. Please
> open a ticket along the lines of "allow LABEL_MAKER in component names OR
> document that it should not be used."
>
> On Tue, May 9, 2017 at 7:33 PM, Manoj Samel
> wrote:
>
> > I
Filed https://issues.apache.org/jira/browse/SLIDER-1227. Thanks for
everyones time !
On Wed, May 10, 2017 at 12:17 PM, Manoj Samel
wrote:
> Thanks Billie for role group explanation, seems like a good feature to
> have !
>
> Thinking a bit about the role group, following are
CDH 5.5.1 cluster with Kerberos, slider version 0.80
Sometimes Slider commands start hanging
slider list --containers
[r...@s-76zyl02.sys.az1.eng.pdx.wd ~]# slider list spas --containers
2017-09-18 21:44:45,659 [main] INFO tools.SliderUtils - JVM initialized
into secure mode with kerberos real
Any thoughts ?
On Mon, Sep 18, 2017 at 3:22 PM, Manoj Samel
wrote:
>
> CDH 5.5.1 cluster with Kerberos, slider version 0.80
>
> Sometimes Slider commands start hanging
>
> slider list --containers
>
> [r...@s-76zyl02.sys.az1.eng.pdx.wd ~]# slider list spas --conta
the RM UI and load the ApplicationMaster web ui for
> this app?
>
> -Gour
>
> On 9/21/17, 11:00 AM, "Manoj Samel" wrote:
>
> >Any thoughts ?
> >
> >On Mon, Sep 18, 2017 at 3:22 PM, Manoj Samel
> >wrote:
> >
> >>
> >>
fails. Seems like your
> issue is intermittent. RPC timeout for CLIs are set to 15 secs, so there
> could be several reasons for which the timeout occurs. Do you see any
> network/routing issue to connect to the host where the AM is running?
>
> -Gour
>
> On 9/21/17, 12:31 PM, &qu
Hello,
Slider version 0.80 on CDH 5.5.1 cluster with kerberos
Slider upgrade --template /xxx/appConfig.json --resources
/xxx/resources.json --queue tenant --force failed with following trace
2018-01-17 20:31:23,030 [main] INFO tools.SliderUtils - JVM initialized
into secure mode with kerberos
/.slider/cluster/spas/readlock
>
>
> -Gour
>
> On 1/17/18, 1:05 PM, "Manoj Samel" wrote:
>
> >Hello,
> >
> >Slider version 0.80 on CDH 5.5.1 cluster with kerberos
> >
> >Slider upgrade --template /xxx/appConfig.json --resources
> >/xxx/reso
Hello,
Slider version 0.80 with CDH 5.5.1
Investigating a instance where Slider application errored out.
slider.agent.log for many components show following trace - noticed that
the "hostname" key is actually the container name e.g. "hostname":
"container_e14_1513412386901_898934_03_03___a
David,
When local disks on the host running node manager are more than 90% full,
nodemanager gives message like "10/12 local-dirs are bad:". In such cases,
the node manager service keeps running but is not servicing any
applications.
Check if the host had multiple disk more than 90% full.
Hope
Hello,
First of all, apologies if this forum is for slider developers and not end
users. The slider user forum hosted @ Hortonworks community does not seem
to have any significant traffic so cross-posting here
Please take a look @
http://hortonworks.com/community/forums/topic/slider-application-d
:
> Slider does not provide REST APIs to create/manage applications yet. Until
> then a Java client application can bundle the slider-core jar and use the
> Java APIs provided in SliderClien.java.
>
>
> -Gour
>
> On 11/3/15, 5:12 PM, "Manoj Samel" wrote:
>
>
of #1
Is my understanding correct?
Thanks,
On Wed, Nov 4, 2015 at 10:56 AM, Steve Loughran
wrote:
>
> On 4 Nov 2015, at 01:53, Manoj Samel manojsamelt...@gmail.com>> wrote:
>
> Thanks Gour,
>
> Is there a plan to provide REST APIs in near future ?
>
> In use case
Hi,
Cross posting in Dev since no replies on forum. Any feedback will be
appreciated.
http://hortonworks.com/community/forums/topic/hdp-dependency-for-secured-cluster/
Thanks,
Thanks Steve !
Manoj
On Fri, Nov 6, 2015 at 3:44 PM, Steve Loughran
wrote:
>
> > On 6 Nov 2015, at 18:31, Manoj Samel wrote:
> >
> > Hi,
> >
> > Cross posting in Dev since no replies on forum. Any feedback will be
> > appreciated.
> >
> >
Hi,
I deployed the sample MEMCACHED application using slider without any issue.
Then I tried to add a new component "MEMCACHED_2" by adding it in 3 files
and running "slider upgrade" command
metainfo.xml
MEMCACHED
MASTER
Never mind .. found out the issue (metainfo update with comp name was not
packaged)
On Mon, Nov 30, 2015 at 12:47 PM, Manoj Samel
wrote:
> Hi,
>
> I deployed the sample MEMCACHED application using slider without any issue.
>
> Then I tried to add a new component "MEMCACHED_
Hi,
Looking at the slider example app memcached README.txt, it seems that it is
required to package metainfo.xml when creating the jmemcached-1.0.0.zip
file. It is not required to package appConfig.json or resources.json (even
the appConfig-default.json or resourses-default.json), as these are giv
!!!
On Mon, Nov 30, 2015 at 6:12 PM, Manoj Samel
wrote:
> Hi,
>
> Looking at the slider example app memcached README.txt, it seems that it
> is required to package metainfo.xml when creating the jmemcached-1.0.0.zip
> file. It is not required to package appConfig.json or resources.j
e": {
"type": "tarball",
"name": "files/jmemcached-1.0.0.tar"
}
}
}
}
}
}
}
Any help will be appreciated ...
Thanks,
On Mon, Nov 30, 2015 at 6:43 PM, Manoj Samel
wrote:
> As a fo
n trying to read
the metainfo.
What are the mandatory entries in metainfo.json when supplied with
--metainfo ?
Any thoughts on this will be appreciated.
On Mon, Nov 30, 2015 at 7:01 PM, Manoj Samel
wrote:
> I tried slider create <> --metainfo metainfo.json and other flags. It gave
>
Can some explain the documented limitation in
http://slider.incubator.apache.org/docs/slider_specs/specifying_exports.html
In the section on exporting data at component level,
"Currently, only the host name and dynamically allocated port values can be
exported"
Why only host/port exports are all
Hi,
Setup - slider version .80, secure hadoop cluster, registry enabled with
security
Noticed using *** zookeeper client (zkCli.sh) *** that when slider
application is stopped, the registry entries in
zookeeper /registry/users/xxx/services/org-apache-slider/abc/components do
not get deleted. Howe
Dec 23, 2015 at 11:12 AM, Jon Maron wrote:
>
> > On Dec 23, 2015, at 2:03 PM, Manoj Samel
> wrote:
> >
> > Hi,
> >
> > Setup - slider version .80, secure hadoop cluster, registry enabled with
> > security
> >
> > Noticed using *** zookeep
Hi,
Setup - Slider 0.80, secure cluster, registry enabled with secure
I am trying to export per component host/slider allocated port using
following
metainfo.xml
port2
MEMCACHED port2 service deployed by Slider
1.0.0
YARN-APP
None
;
> > On 23 Dec 2015, at 20:21, Jon Maron wrote:
> >
> >>
> >> On Dec 23, 2015, at 3:13 PM, Manoj Samel
> wrote:
> >>
> >> Thanks Jon,
> >>
> >> I thought the ZK registry was to reflect the live information about the
> >>
Hi,
I have been exploring slider for some time and would like to contribute to
make documentation better to improve the ramp up time & effort for new
slider users.
Can someone point how do I contribute to doc updates ?
Thanks,
Manoj
gt; >
> >> On Dec 23, 2015, at 6:27 PM, Manoj Samel
> wrote:
> >>
> >> Jon, as to your last comment -
> >>
> >> As I understand, the idea of pushing slider registry to more general
> yarn
> >> registry was to make it easier for clien
Hi,
Slider version .80 on secured cluster.
I am seeing the pattern consistently
1. Create cluster with just slider AM. Works fine
2. Slider upgrade to add 1 component. Works fine
3. Slider upgrade to add 2nd component. After upgrade, the application goes
in FAILED state.
I start it manually usin
Hi,
Slider 0.80 with secured cluster.
Say I have cluster with 1 component running. I do upgrade and add second
component. After running the upgrade command, when I do "slider list xxx
--containers" the first running component shows "awaiting heartbeat" for
like 2 minutes ... There is also pause i
On Mon, Jan 4, 2016 at 4:51 PM, Manoj Samel
wrote:
> Hi,
>
> Slider version .80 on secured cluster.
>
> I am seeing the pattern consistently
>
> 1. Create cluster with just slider AM. Works fine
> 2. Slider upgrade to add 1 component. Works fine
> 3. Slider upgrade
Hi,
Use case needs adding and removing components dynamically.
What is the proper way to remove a component ? It seems it can be done
either by doing a "slider flex" with component count set to 0 or by doing a
"slider upgrade" where the component is deleted from metainfo.xml,
appconfig.json and r
e a good
feature to be able to remove components permanently for use cases like
above; where many users will come and go. If you think this is a valid
feature ask, I can file a Jira
Thanks,
Manoj
On Tue, Jan 5, 2016 at 2:42 AM, Steve Loughran
wrote:
>
> > On 5 Jan 2016, at 02:16, Man
one by setting yarn.component.instances for
user1 to 0 ( I am sure flex 0 would do same ) to achive stopping component.
I understand this history will be required if someone did a flex 0 and then
need to do flex N (N > 0) later. However, in the use cases where this
component will never be used a
Loughran
wrote:
>
> > On 5 Jan 2016, at 00:51, Manoj Samel wrote:
> >
> > Hi,
> >
> > Slider version .80 on secured cluster.
> >
> > I am seeing the pattern consistently
> >
> > 1. Create cluster with just slider AM. Works fine
> >
Thanks Steve !
Filed https://issues.apache.org/jira/browse/SLIDER-1046
On Wed, Jan 6, 2016 at 7:03 AM, Steve Loughran
wrote:
>
> > On 5 Jan 2016, at 20:34, Manoj Samel wrote:
> >
> > For delete case,
> >
> > Digging further into slider HDFS data in histor
Slider 0.80 on secured cluster. Each component exports its host/port and
some other info
I am looking @ exported values at
http://host:1025/ws/v1/slider/publisher/slider/componentinstancedata
1 - Just export host/port - This works
my_info
${THIS_H
I can add
any constant or export multiple ports - so not clear what is the reason
behind blocking publishing of any non-number info ...
Should I file a Jira for this ?
Thanks
Manoj
On Wed, Jan 6, 2016 at 11:30 AM, Manoj Samel
wrote:
> Slider 0.80 on secured cluster. Each component exports
Any thoughts ?
Thanks,
On Thu, Jan 7, 2016 at 10:25 AM, Manoj Samel
wrote:
> Hi,
>
> Slider 0.80 with secured cluster. Registry is enabled *** without ***
> security
>
> Export is configured as follows
>
> 1. No application level export
> 2. Component level port
gt; Hey Manoj,
>
> I'm not sure at a glance, but it might be a bug in Slider's use of the
> registry? I'm not quite sure, I'd have to dig into the source to be sure.
>
> Do you have an example that I could run which exhibits this?
>
>
> Manoj Samel wrote:
&
Slider version 0.80 with secured cluster
Use case is to create a component reflecting user name. It seems the only
valid character in component name besides [A-Z][a-z[0-9] is underscore '_'.
Attempt to create a component with characters like dash '-' or many other
characters fail to bring up the
Jon,
I replaced with actual component name.
Thanks,
On Tue, Jan 12, 2016 at 11:43 AM, Jon Maron wrote:
> Did you replace the actual comp name with , or do you actually
> have the ‘<‘ and ‘>’ characters in the name?
>
> > On Jan 12, 2016, at 2:40 PM, Manoj Samel
&g
The component names are currently expected to follow a
> naming convention that allows for DNS compatible names, and dashes are
> included in that character set. The fact that the endpoint did not appear
> may be related to some other issue. The AM logs may help here as well.
>
>
ocally.
>
> On Jan 12, 2016, at 4:27 PM, Manoj Samel manojsamelt...@gmail.com>> wrote:
>
> Hi Jon,
>
> I tried the OOB jmemcached with component name my-name. You may be able to
> reproduce by trying the jmemcached with component name my-name
>
> Attached are the slide
,"empty":true},"componentinstancedata":{"description":"ComponentInstanceData","updated":0,"entries":{},"empty":true}}}
Thanks,
Manoj
On Tue, Jan 12, 2016 at 3:03 PM, Gour Saha wrote:
> When you do the below get, you
ts and
> resource paths to AM) using which you can retrieve the actual data. Most
> of the information is stored in AM runtime memory and not persisted
> anywhere. In case of AM failure all the AM state is rebuilt.
>
> -Gour
>
> On 1/12/16, 5:24 PM, "Manoj Samel" wrote:
&
Filed https://issues.apache.org/jira/browse/SLIDER-1056
Thanks,
Manoj
On Tue, Jan 12, 2016 at 5:45 PM, Gour Saha wrote:
> Please file a jira.
>
> On 1/6/16, 12:20 PM, "Manoj Samel" wrote:
>
> >Dug further into this ... Appears that only the variables that
nformation is stored in AM runtime memory and not persisted
> >> anywhere. In case of AM failure all the AM state is rebuilt.
> >>
> >> -Gour
> >>
> >>> On 1/12/16, 5:24 PM, "Manoj Samel" wrote:
> >>>
> >>> Hi Gour,
>
Hi,
In recent thread about existing components taking long time to heartbeat
after upgrade; it was mentioned that "slider AM restarts itself to pick up
Yarn parameters"
If the upgrade simply consist of starting a new component, is the AM
restart still needed ? I.e. if the hadoop cluster configura
ould be stopped for good. This is achieved at present by naming each
component after respective user and stopping that when user is gone.
Thanks,
Manoj
On Wed, Jan 13, 2016 at 10:11 AM, Steve Loughran
wrote:
> you should just be flexing here
>
> > On 13 Jan 2016, at 09:19, Man
;m with you Manoj. It seems wrong to not see the exports for
> >jmemcached.
> >
> >Gour -- am I missing something too?
> >
> >Manoj Samel wrote:
> >> Shouldn't it also show component host:port in ZK when they are being
> >> exported in registr
Hi,
Any update on this ? Did you get a chance to try the memcache with the
component name I used above ?
Thanks,
Manoj
On Tue, Jan 12, 2016 at 4:37 PM, Steve Loughran
wrote:
>
> > On 12 Jan 2016, at 12:14, Jon Maron wrote:
> >
> > OK. So was there a ‘-‘ or some other character in your compo
ot; wrote:
>
> >Sorry - got pulled off on some other tasks. I¹ll try to take a look over
> >the next day or so.
> >
> >> On Jan 14, 2016, at 11:31 AM, Manoj Samel
> >>wrote:
> >>
> >> Hi,
> >>
> >> Any update on thi
t name?
>
> -Gour
>
> On 1/14/16, 9:42 AM, "Manoj Samel" wrote:
>
> >Hi Gour,
> >
> >1. The issue is with component name. The code you pointed is for cluster
> >name
> >2. As mentioned above, the component name does not seem to allow dash
oach. The overhead of a Slider AM is very low.
>
> -Gour
>
> On 1/13/16, 10:35 AM, "Manoj Samel" wrote:
>
> >Flex isn't option for couple of reasons in above use case
> >
> >1. For each new user, different parameters need to be passed when
> &
Any thoughts ? I feel this is a bug - should I file Jira to track ?
Thanks,
On Wed, Jan 13, 2016 at 1:35 PM, Manoj Samel
wrote:
> At some point, it is expected that the component host:port info ( i.e.
> component discovery) is available in registry - Right ?
>
> On Wed, Jan 13, 2
I have filed https://issues.apache.org/jira/browse/SLIDER-1064
Thanks,
Manoj
On Fri, Jan 15, 2016 at 9:48 AM, Manoj Samel
wrote:
> Any thoughts ? I feel this is a bug - should I file Jira to track ?
>
> Thanks,
>
> On Wed, Jan 13, 2016 at 1:35 PM, Manoj Samel
> wrote:
>
No problem. Please do take a look & comment when you get a chance though.
Thanks,
Manoj
On Mon, Jan 18, 2016 at 12:54 PM, Gour Saha wrote:
> Thanks for filing the jira. I think that¹s the best way to capture it.
>
> -Gour
>
> On 1/18/16, 11:48 AM, "Manoj Samel" w
sible ? Should I file a feature jira ?
Thanks
On Mon, Jan 18, 2016 at 3:05 PM, Gour Saha wrote:
> Inline.
>
> On 1/14/16, 4:52 PM, "Manoj Samel" wrote:
>
> >Hi Gour,
> >
> >The use case is described in
> >https://issues.apache.org/jira/browse
h. Am I missing something in the code :(
>
> -Gour
>
> On 1/14/16, 10:33 AM, "Manoj Samel" wrote:
>
> >@Gour - yes, it does work when component name does not have a dash. It
> >also
> >works when component name has underscore(s)
> >
> >Thanks,
&g
Improvement jira filed - https://issues.apache.org/jira/browse/SLIDER-1066
Thanks all for your time and insights.
Manoj
On Mon, Jan 18, 2016 at 6:56 PM, Gour Saha wrote:
> Yes please file a jira. Let’s discuss it there further.
>
> On 1/18/16, 3:57 PM, "Manoj Samel" wro
Hi,
The latest available release 0.81 has a blocker
https://issues.apache.org/jira/browse/SLIDER-1010 as far as secured cluster
is involved.
The fix looks non-intrusive and of minimal risk. Would it possible to
backport it to 0.81 ?
Thanks,
Manoj
Awesome !!!
Filed https://issues.apache.org/jira/browse/SLIDER-1067 ... Seems I don't
have permission to assign the Jira. Steve - could you please assign
yourself ?
Thanks,
On Wed, Jan 20, 2016 at 11:14 AM, Steve Loughran (JIRA)
wrote:
>
> [
> https://issues.apache.org/jira/browse/SLIDER
Hello,
Any update on this ? Should I file a Jira ?
Thanks,
On Tue, Jan 19, 2016 at 8:58 AM, Manoj Samel
wrote:
> Hi Jon,
>
> Did you get a chance to try to reproduce the issue with component
> containing a dash ?
>
> Thanks,
>
> Manoj
>
> On Thu, Jan 14, 2016
2016, at 12:15 PM, Manoj Samel
> wrote:
> >
> > Hello,
> >
> > Any update on this ? Should I file a Jira ?
> >
> > Thanks,
> >
> > On Tue, Jan 19, 2016 at 8:58 AM, Manoj Samel
> > wrote:
> >
> >> Hi Jon,
> >
Hi,
Slider 0.80 running on CDH 5.5.1 secured cluster fails to create slider-AM
with error "User is not based on a keytab in a secure deployment".
1) I had same slider version working without issue on CDH 5.4.2 without any
issues.
2) I thought the issue was in slider 0.81 (
https://issues.apache.o
Any help on this will be appreciated !!!
Thanks,
Manoj
On Mon, Feb 8, 2016 at 5:54 PM, Manoj Samel
wrote:
> Hi,
>
> Slider 0.80 running on CDH 5.5.1 secured cluster fails to create slider-AM
> with error "User is not based on a keytab in a secure deployment".
>
>
in cdh 5.5.1 ?
2. I had opened SLIDER-1067 for backport of slider fix to .82 release. Any
idea when 0.82 will be available ?
Thanks,
On Wed, Feb 10, 2016 at 3:45 AM, Steve Loughran
wrote:
>
> > On 9 Feb 2016, at 01:54, Manoj Samel wrote:
> >
> > Hi,
> >
> >
Hello,
Environment is slider .80 on Hadoop 2.6 secured cluster
A component is launched for each distinct user of the service (via
upgrade). E.g. when user A accesses service, do a "upgrade" and create a
component for user A. When user B comes, create another component for user
B etc.
At present,
Any thoughts ? Feedback ?
On Fri, Apr 8, 2016 at 10:30 AM, Manoj Samel
wrote:
> Hello,
>
> Environment is slider .80 on Hadoop 2.6 secured cluster
>
> A component is launched for each distinct user of the service (via
> upgrade). E.g. when user A accesses service, do a &quo
quot;B" and does a "exec" for the java component
On Fri, Apr 8, 2016 at 10:30 AM, Manoj Samel
wrote:
> Hello,
>
> Environment is slider .80 on Hadoop 2.6 secured cluster
>
> A component is launched for each distinct user of the service (via
> upgrade). E.g. whe
user "B")
Any help in identifying why is the parent process dying ?
Thanks in advance !!
Manoj
PS : Please ignore my last mail sent with same title few minutes back. I
hit return by mistake when it was not complete :(
On Fri, Apr 8, 2016 at 10:30 AM, Manoj Samel
wrote:
> Hello,
&g
nt process
which is "/bin/bash --login ..." and which launches a subprocess as
specified in execute(). Can someone confirm this ?
On Thu, Apr 21, 2016 at 2:40 PM, Manoj Samel
wrote:
> Hi,
>
> See use case background below
>
> I have implemented option 2 mentioned below
Thanks,
Manoj
------ Forwarded message --
From: Manoj Samel
Date: Thu, Apr 21, 2016 at 2:40 PM
Subject: Need Help !: Run each component of application as different user
To: dev@slider.incubator.apache.org
Hi,
See use case background below
I have implemented option 2 mentioned below (as
ss, like the system command would. Perhaps if you used the system
> command in your C code, that would produce a different result.
>
> Billie
>
> On Fri, Apr 22, 2016 at 12:17 PM, Manoj Samel
> wrote:
>
> > Hello Again !
> >
> > One more observation .. hopeful
security in parity with what
the traditional map-reduce does. I believe this will be important feature
as slider gets more adoption for running custom services ( beyond Hbase
etc. which could be run as single user)
Thanks,
Manoj
On Wed, May 4, 2016 at 7:28 AM, Josh Elser wrote:
> Manoj Samel wr
Hello,
When using ALLOCATED_PORT clause, there is a option "PER_CONTAINER".
Can someone explain what does "PER_CONTAINER" option does ? It says keep
port allocation private to container. What does that means ? If multiple
containers are chosen to on same host machine, will this cause issue ?
Whe
y
> another application on the same box.
>
> Is your cluster idle or do you have other slider apps running? Do you have
> more complete output of the logs and possibly the appConfig that you can
> share? Are you sure it's the AM failing to start and not a service within
ks.AbstractQueuedSynchronizer$ConditionObject.reportInterruptAfterWait(AbstractQueuedSynchronizer.java:2014)
at
java.util.concurrent.locks.AbstractQueuedSynchronizer$ConditionObject.await(AbstractQueuedSynchronizer.java:2048)
at
java.util.concurrent.LinkedBlockingQueue.take(LinkedBlockingQueue.java:4
nteger.parseInt(m.group()));
} else {
m = NUMBER_RANGE.matcher(range.trim());
if (m.find()) {
} // else is missing . Add with a exception ???
Thoughts ?
Manoj
On Mon, May 16, 2016 at 6:45 PM, Manoj Samel
wrote:
> Here is slider.log for Slider AM. Note the port r
Filed https://issues.apache.org/jira/browse/SLIDER-1124
On Thu, May 19, 2016 at 7:18 AM, Billie Rinaldi
wrote:
> I agree, it would be good to throw an Exception if the format of the port
> range is bad.
>
> On Wed, May 18, 2016 at 7:15 PM, Manoj Samel
> wrote:
>
> &
Hi,
At present the allowed.ports is applied to AM as well as containers
While this may be a valid use case for certain deployments, there are other
use cases
1. The "allowed.ports" should be only applied to certain port(s) for
components that are marked as {ALLOCATED_PORT}{PER_CONTAINER} etc.. T
Hello,
Just wanted to confirm my understanding of PER_CONTAINER for allocated
ports.
Looking @ AgentProviderService.java -
if (!value.contains(PER_CONTAINER_TAG)) {
// If the config property is shared then pass on the already
allocated value
// from any container
If
Any thoughts ?
On Wed, May 25, 2016 at 11:27 AM, Manoj Samel
wrote:
> Hi,
>
> At present the allowed.ports is applied to AM as well as containers
>
> While this may be a valid use case for certain deployments, there are
> other use cases
>
> 1. The "allowed.por
assigned ports from the specified ranges.
>
> > On Jun 1, 2016, at 12:15 PM, Manoj Samel
> wrote:
> >
> > Any thoughts ?
> >
> > On Wed, May 25, 2016 at 11:27 AM, Manoj Samel
> > wrote:
> >
> >> Hi,
> >>
> >> At present th
Hello,
https://issues.apache.org/jira/browse/SLIDER-1114
Trying to implement a new feature where each component of a application is
run as different linux user. I have so far implemented a way to start each
component as different user and accurately keep track status of each
component. The implem
Setup
- Hadoop 2.6 with RM HA, Kerberos enabled
- Slider 0.80
- In my slider-client.xml, I have added all RM HA properties, including the
ones mentioned in http://markmail.org/message/wnhpp2zn6ixo65e3.
Following is the issue
* rm1 is active, rm2 is standby
* deploy and start slider application,
again. Slider AM still keeps running
So, it seems if it starts with rm1 active, then the AM goes to "ACCEPTED"
state when RM fails to rm2. If it starts with rm2 active, then it runs fine
with any switches between rm1 and rm2.
Any feedback ?
Thanks,
Manoj
On Mon, Jul 25, 2016 at 12:25
Hello,
Slider version is 0.80, Hadoop is 2.6 with Kerberos
Slider-client.xml allows specification of full path of hadoop conf
using HADOOP_CONF_DIR. In our case, full hadoop configuration, including
all HA configurations are available in the HADOOP_CONF_DIR for hdfs-site,
core-site and yarn-site.
ors?
>
> -Gour
>
> On 7/25/16, 2:28 PM, "Manoj Samel" wrote:
>
> >Another observation (whatever it is worth)
> >
> >If slider app is created and started when rm2 was active, then it seems to
> >survive switches between rm2 and rm1 (and back). I.e
>
r everything out from
> slider-client.xml.
>
> On 7/25/16, 4:12 PM, "Manoj Samel" wrote:
>
> >Hi Gour,
> >
> >Thanks for your prompt reply.
> >
> >FYI, issue happens when I create slider app when rm1 is active and when
> >rm1
> >fails
sues you are facing, you have to provide additional
> logs for us to understand better. Let¹s start with -
> 1. RM logs (specifically between the time when rm1->rm2 failover is
> simulated)
> 2. Slider App logs
>
> -Gour
>
> On 7/25/16, 5:16 PM, "Manoj Samel"
Filed https://issues.apache.org/jira/browse/SLIDER-1158 with logs and my
analysis of logs.
On Tue, Jul 26, 2016 at 10:36 AM, Gour Saha wrote:
> Please file a JIRA and upload the logs to it.
>
> On 7/26/16, 10:21 AM, "Manoj Samel" wrote:
>
> >Hi Gour,
> >
>
1 - 100 of 196 matches
Mail list logo