Mike_Phelps

updated

10 years ago

M

Mike_Phelps

1 Rookie

•

12 Posts

0

2964

December 2nd, 2016 11:00

ECS 3.0 Community Edition Web UI

I have a four node ECS 3.0 cluster that has been working fine for three weeks.  Recently, attempting to log into the Web UI from all four nodes returns the same error message:


Unable to connect to the service. Service(s) are unavailable, try again later.


S3 connectivity has been intermittent for a time during this UI issue, but it seemed to recover and has been continuing to service application needs. And since this login problem occurs against each of the four nodes, I question whether there's a single log to glean more information from.  ECS service log locations / /opt/emc/caspian/fabric/agent/services/object/main/log

Thoughts?

  • JasonCwik

    1 Rookie

    •

    281 Posts

    2191

    0

    Posted December 7th, 2016 11:00

    I talked to Mike on the phone.  The issue what that the auth service was blocked on all 4 nodes.  We restarted it with:

    sudo kill `pidof authsvc`

    And logins now work to the portal.  We'll work on the cyberduck issue as next.

  • benschumacher

    1 Rookie

    •

    75 Posts

    2191

    0

    Posted December 2nd, 2016 14:00

    Can you goto http:// :9101/ and click the Show DT Initialization Status link in third column midway down?  What's the result of that status call?

    Ben

  • Mike_Phelps

    1 Rookie

    •

    12 Posts

    2191

    0

    Posted December 2nd, 2016 14:00

    thanks for the response Ben Schumacher 

     

    I entered the public ip:9101 and clicked the link you mentioned.  That link changed to the private interface so I changed the url to the /stats/dt/DTInitStat/. Not sure if this was the expected behavior.

     

    some XML:

     

     

     

    512

     

    0

    0

  • 2191

    0

    Posted December 5th, 2016 13:00

    Hi Mike!

    What happens when you try accessing one of the API endpoints?  Same error or something different?

    Thanks!

  • Mike_Phelps

    1 Rookie

    •

    12 Posts

    2191

    0

    Posted December 7th, 2016 07:00

    leveraging Cyberduck, my colleague couldn't access S3 buckets late yesterday. However, he can connect this morning.  The Web UI is still throwing the error posted initially.

  • Mike_Phelps

    1 Rookie

    •

    12 Posts

    2191

    0

    Posted December 8th, 2016 20:00

    now that the Web UI is responsive, and I haven't seen any issues with S3 Browser, I'll close as answered/resolved.  Looking into Local Traffic Manager / Global Traffic Manager to improve on the current configuration in front of ECS Free and Frictionless to supported appliance based deployment.

     

    Thanks Jason Cwik  for jumping on a WebEx yesterday.

     

    Mike