Hi, @Haripriya It looks like there are some problems in your master flags.

> I'm attaching a snapshot of the error I've seen in Chrome with this
email. It'll be great if you can suggest if I'm missing any configuration
or if its some bug.
According to the screenshot you attached, the hostnames are incorrect on
your servers. Mesos WebUI depends on that to find the leading master.
A workaround is to specific the `--hostname` flag when starting your
masters. For example, launch your masters with

```
$ mesos-master --hostname=socrates-nid000xxx.us.cray.com xxx
```

> Is it something to do with a stale state of mesos anywhere or the way I'm
passing cluster? I have a config file named cluster in /etc/mesos-master/
and when I restart the cluster it picks up the config files.

You need to ensure the flags of every master contains
`--cluster=your_cluster_name`.

Could you perform `ps aux |grep mesos-master` on every master and paste
their outputs here?


On Wed, Dec 7, 2016 at 4:39 AM, Haripriya Ayyalasomayajula <
aharipriy...@gmail.com> wrote:

> Hello, @Haosdent,
>
> Thanks for suggesting these.
> I'm attaching a snapshot of the error I've seen in Chrome with this email.
> It'll be great if you can suggest if I'm missing any configuration or if
> its some bug.
>
> And for the second part, my `/master/state` end point does not return
> "cluster" anywhere. It returned 75k lines of json so I'm not pasting all of
> it.
> {
>     "activated_slaves": 37.0,
>     "build_date": "2016-11-16 01:31:49",
>     "build_time": 1479259909.0,
>     "build_user": "centos",
>     "completed_frameworks": [
>         {
>             "active": true,
>   ..........
>
>
>
>     "start_time": 1480967418.42687,
>     "unregistered_frameworks": [],
>     "version": "1.1.0"
> }
>
> Is it something to do with a stale state of mesos anywhere or the way I'm
> passing cluster? I have a config file named cluster in /etc/mesos-master/
> and when I restart the cluster it picks up the config files.
>
> On Mon, Dec 5, 2016 at 6:24 PM, haosdent <haosd...@gmail.com> wrote:
>
>> Hi, @Haripriya
>>
>> > (less than 1 min though the  jobs are running just fine).
>> > Is there any new configuration that has to be added?
>>
>> We change to use JSONP to send requests in WebUI since 1.0 May I have
>> your error log in Safari, Chrome and Firefox?
>> You could open it via https://developers.google.
>> com/web/tools/chrome-devtools/console/
>>
>> > The UI does not display the name of the cluster despite using the
>> --cluster flag.
>> --cluster flag works fine for me. May you paste your `/master/state`
>> endpoint at the email, I would like to check the value of `cluster` field
>> in it.
>>
>> On Tue, Dec 6, 2016 at 5:34 AM, Haripriya Ayyalasomayajula <
>> aharipriy...@gmail.com> wrote:
>>
>>> Hi all,
>>>
>>> I have two issues with the web UI in Mesos 1.1
>>>
>>> 1.
>>>
>>> Earlier when I was using Mesos 0.28, mesos web UI would try to reconnect
>>> only when there are network issues or when there is a newly elected leader.
>>> After upgrade to 1.1, we see that it won't work (shows no leader is elected
>>> even when there is a leader elected and jobs are running happily ) on
>>> safari, works on chrome and firefox but tries to re-connect very often
>>> (less than 1 min though the  jobs are running just fine).
>>>
>>> Is there any new configuration that has to be added?
>>>
>>>
>>> 2. The UI does not display the name of the cluster despite using the
>>> --cluster flag.
>>>
>>> /usr/sbin/mesos-master --zk=zk://mesos1:2181,mesos2:2181,mesos3:2181/
>>> mesos --port=5050 --log_dir=/var/log/mesos --acls=/etc/mesos_acls.json
>>> --authenticate_frameworks=true --cluster="cluster1"
>>> --credentials=/etc/auth/credentials --quorum=2 --work_dir=/var/lib/mesos
>>>
>>>
>>> I also tried adding the name of the cluster without quotes: cluster1
>>> instead of "cluster1", but that doesn't work either.
>>>
>>> /usr/sbin/mesos-master --zk=zk://mesos1:2181,mesos2:2181,mesos3:2181/
>>> mesos --port=5050 --log_dir=/var/log/mesos --acls=/etc/mesos_acls.json
>>> --authenticate_frameworks=true --cluster=cluster1
>>> --credentials=/etc/auth/credentials --quorum=2 --work_dir=/var/lib/mesos
>>> I greatly appreciate any help!
>>>
>>> --
>>> Thanks,
>>> Haripriya
>>>
>>
>>
>>
>> --
>> Best Regards,
>> Haosdent Huang
>>
>
>
>
> --
> Thanks,
> Haripriya
>
>


-- 
Best Regards,
Haosdent Huang

Reply via email to