Trove show retry too many times and then hit 504 Gateway Timeout

[Fan Zhang]

When I use command 'trove show <instanceId>' to get details of a normal existent instance, everything goes fine. When I try to show an inexistent instance, I get 'ERROR: No instance with a name or ID of '3e724b0e-ad35-4a01-aba0-bcb7baf0918x' exists.' It looks OK, either.

But when I have lots of trove instances, some of them are ACTIVE while their nova instances may be SHUTDOWN. Using trove show to show an inexsitent instance, I see there are lots of request to list trove instances and then get some instances which are not what I appoint, if its nova instance is SHUTDOWN, there is great chance to encounter 504 Gateway Timeout.

Here is the log [1]

I look through python-troveclient codes [2], I think maybe it retryes too many times to find an instance.

[1] http://paste.ubuntu.com/25807496
[2]https://github.com/openstack/python-troveclient/blob/565a6e0a5b9467df9880ed4c831ff0bca1a5084a/troveclient/utils.py#L210

So I open this bug, hope someone can give some feedback.

[Yao Wang (nihaome521)]

I tried several times, but failed to replicate this bug.

[root@oracle12c /etc/trove]# nova list|grep rdswy
| b4d6ace6-887a-4391-a238-582c433349ce | rdswy | SHUTOFF | - | Shutdown | private-share-v1249_10.132.49=10.132.49.149; bigdata_network=10.132.53.30 |
[root@oracle12c /etc/trove]#
[root@oracle12c /etc/trove]# trove list
+--------------------------------------+-------+-----------+-------------------+--------+-----------+------+--------+
| ID | Name | Datastore | Datastore Version | Status | Flavor ID | Size | Region |
+--------------------------------------+-------+-----------+-------------------+--------+-----------+------+--------+
| d884ba96-a6bf-4e62-adb0-1e93a8c8e80f | rdswy | mysql | 5.6-cluster | ACTIVE | 310 | 1 | |
+--------------------------------------+-------+-----------+-------------------+--------+-----------+------+--------+
[root@oracle12c /etc/trove]# trove show d884ba96-a6bf-4e62-adb0-1e93a8c8e80f
+-------------------+---------------------------------------------------------------------------------------------+
| Property | Value |
+-------------------+---------------------------------------------------------------------------------------------+
| created | 2017-11-07T16:46:22 |
| datastore | mysql |
| datastore_version | 5.6-cluster |
| flavor | 310 |
| id | d884ba96-a6bf-4e62-adb0-1e93a8c8e80f |
| ip | 10.132.49.149 |
| name | rdswy |
| networks | {u'name': u'private-share-v1249_10.132.49', u'id': u'141ed22d-36ba-4436-af79-b05bbf9016e0'} |
| status | ACTIVE |
| updated | 2017-11-07T16:46:25 |
| volume | 1 |
+-------------------+---------------------------------------------------------------------------------------------+
[root@oracle12c /etc/trove]# trove show d884ba96-a6bf-4e62-adb0-1e93a8c8e80g
ERROR: No instance with a name or ID of 'd884ba96-a6bf-4e62-adb0-1e93a8c8e80g' exists.
[root@oracle12c /etc/trove]#

[Zhao Chao (zhaochao1984)]

Need more information for this, and for a "504 Gateway Timeout" error, it is likely caused by some load balancer layer.

Change to "Won't Fix" first, you can reopen this if more information is provided.