perfSONAR User Q&A and Other Discussion

["Robinson, John-Paul" <>]

More evidence that esmond is happy with cassandra:

workshop@soxws14:/usr/lib/esmond$ sudo /usr/sbin/esmond_manage check
cassandra_db [INFO] Checking/creating column families
cassandra_db [INFO] Schema check done
cassandra_db [DEBUG] Opening ConnectionPool
cassandra_db [INFO] Connected to ['localhost:9160']
cassandra_db [INFO] Checking/creating column families
cassandra_db [INFO] Checking/creating column families
cassandra_db [INFO] Schema check done
cassandra_db [INFO] Schema check done
cassandra_db [DEBUG] Opening ConnectionPool
cassandra_db [DEBUG] Opening ConnectionPool
cassandra_db [INFO] Connected to ['localhost:9160']
cassandra_db [INFO] Connected to ['localhost:9160']
System check identified no issues (0 silenced).

I'm focusing on the esmond permissions from clients now.

sudo /usr/sbin/esmond_manage help add_user_ip_addr

On 09/21/2018 02:29 PM, Robinson, John-Paul wrote:
> Thanks for the hints.  I agree these are the likely problem places.
>
> I stopped cassandra and that cleaned up the processes.   Started it up
> but it has the same status.  While it was down, django did log
> complaints about not getting to cassandra.  It doesn't say anything
> about getting to cassandra, so it either is or is not reaching cassandra. :)
>
> The systemctl status may be due to systemctl loosing track of the
> process it starts.  I'll look there.
>
> Considering the danjgo config, does it need a user/pass to connect
> cassandra?  In the config that go created we have:
>
> cassandra_servers = localhost:9160
> cassandra_user =
> cassandra_pass =
>
> Given the complaints when cassandra went down, I'm favoring the IP authz
> rules issue.  Any hints on what needs to be set here?  Is using a
> generic user name valid or does that need to be explicitly configured?
>
> I do see clients connecting to the maddash server via netstat (https and
> owamp) so I assume they are succeeding on the TCP level at least.
>
> John-Paul
>
>
> On 09/21/2018 02:10 PM, Michael Johnson wrote:
>> Hi,
>>
>> Comments below:
>>
>> On Fri, Sep 21, 2018 at 05:12:37PM +0000, Robinson, John-Paul wrote:
>>
>>> We suspect something is up with cassandra the status shows somewhat
>>> confusing output "active (exited)":
>>>
>>> workshop@soxws14:~/projects/mesh$ sudo systemctl status cassandra
>>> ● cassandra.service - SYSV: Starts and stops Cassandra
>>>     Loaded: loaded (/etc/rc.d/init.d/cassandra; bad; vendor preset:
>>> disabled)
>>>     Active: active (exited) since Fri 2018-09-21 16:43:46 UTC; 24min ago
>>>       Docs: man:systemd-sysv-generator(8)
>>>    Process: 5622 ExecStop=/etc/rc.d/init.d/cassandra stop (code=exited,
>>> status=1/FAILURE)
>>>    Process: 5934 ExecStart=/etc/rc.d/init.d/cassandra start
>>> (code=exited, status=0/SUCCESS)
>>>
>>> Sep 21 16:43:46 soxws14.sox.net systemd[1]: Starting SYSV: Starts and
>>> stops Cassandra...
>>> Sep 21 16:43:46 soxws14.sox.net su[5943]: (to cassandra) root on none
>>> Sep 21 16:43:46 soxws14.sox.net cassandra[5934]: Starting Cassandra: OK
>>> Sep 21 16:43:46 soxws14.sox.net systemd[1]: Started SYSV: Starts and
>>> stops Cassandra.
>> To clarify a couple points, "active (exited)" means the service is
>> *enabled* (this is what "active" means in this context), but it is not
>> running (exited).
>>
>> I would suggest stopping cassandra using
>> $ sudo systemctl stop cassandra
>>
>> and then checking to make sure there are no more cassandra processes
>> running:
>>
>> $ ps aux | grep cassandra
>>
>> If there are any remaining processes, kill them. Then start cassandra
>> back up.
>>
>> Your statement that local tests are being archived leads me to believe
>> that cassandra is running and your other hosts are unable to
>> authenticate (in this case, your ip auth may not be working).
>>
>> On the other hand, those Django errors usually mean Cassandra is
>> down/unreachable from the point of view of esmond, regardless of what
>> cassandra itself tells you.
>> These two ideas do seem to conflict, but I'd try these steps to see
>> where they gets you.
>>
>> Thanks,
>> Michael
>>