[
https://issues.apache.org/jira/browse/AIRAVATA-2376?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15991236#comment-15991236
]
Marcus Christie commented on AIRAVATA-2376:
-------------------------------------------
Just documenting a quick fix I tried to apply but it didn't work but it might
be the start of a complete solution.
The problem is in the LATEST_EXPERIMENT_STATUS view. It orders
EXPERIMENT_STATUS by TIME_OF_STATE_CHANGE but sometimes two statuses have the
same TIME_OF_STATE_CHANGE.
There are a couple of ways to fix this. One way that I tried is that the
TIME_OF_STATE_CHANGE only has accuracy to the nearest second. I tried adding
millisecond accuracy like so
{code:sql}
ALTER TABLE EXPERIMENT_STATUS MODIFY COLUMN TIME_OF_STATE_CHANGE TIMESTAMP(6)
NOT NULL DEFAULT CURRENT_TIMESTAMP(6) ON UPDATE CURRENT_TIMESTAMP(6);
{code}
See also https://mariadb.com/kb/en/mariadb/microseconds-in-mariadb/
This changed the definition of the column but the values stored are still being
put in as a whole second. I think maybe the client code needs to be updated
somehow. According to http://stackoverflow.com/a/25803063 one has to enable
useFractionalSeconds in the JDBC driver.
> Duplicating experiment records in Experiment statistics and Experiment browse
> -----------------------------------------------------------------------------
>
> Key: AIRAVATA-2376
> URL: https://issues.apache.org/jira/browse/AIRAVATA-2376
> Project: Airavata
> Issue Type: Bug
> Components: PGA PHP Web Gateway, Registry API
> Affects Versions: 0.17
> Environment: https://seagrid.org and all other gateways as well
> Reporter: Eroma
> Assignee: Marcus Christie
> Fix For: 0.17
>
>
> When experiments are launched there are two records in EXPERIMENT SUMMARY
> table for LAUNCHED and EXECUTING state which has the exact same state change
> time. In PGA however two records show up for the same experiment and with
> same status. e.g.: two exact records for a single experiment. This only
> appears until the experiment is FAILED or COMPLETED. But both users and
> gateway admins notices this and this needs fixing.
--
This message was sent by Atlassian JIRA
(v6.3.15#6346)
