many retries can overflow the results column in atomdetails

[Greg Hill]

If you have a task that retries a lot of times (up to 80 in this particular case, which I know is excessive, but makes sense in context), it can eventually write a JSON string to the database that is longer than the field can handle. In the default mysql backend, mysql will silently truncate that JSON string, making it no longer valid. Any further retries then fail, because they can't load the history due to it being invalid JSON.

For example (elided for sanity):

Original exception being dropped: ['Traceback (most recent call last):\n', ..File "/usr/lib/python2.7/json/decoder.py", line 382, in raw_decode\n obj, end = self.scan_once(s, idx)\n', 'ValueError: Unterminated string starting at: line 1 column 65497 (char 65497)\n']

[OpenStack Infra (hudson-openstack)]

Reviewed: https://review.openstack.org/153443
Committed: https://git.openstack.org/cgit/openstack/taskflow/commit/?id=1fc1a7e4713c27f537a3f6192bea259a0d63ca64
Submitter: Jenkins
Branch: master

commit 1fc1a7e4713c27f537a3f6192bea259a0d63ca64
Author: Joshua Harlow <email address hidden>
Date: Thu Feb 5 17:46:58 2015 -0800

    Add warning to sqlalchemy backend size limit docs

    Partial-Bug: #1416088

    Change-Id: I21bde00445892a4c734592cbcc143f23085e5660

[Joshua Harlow (harlowja)]

So I also just tried this:

http://paste.openstack.org/show/198192/

The usage of 'mysql_sql_mode': 'TRADITIONAL', there will avoid saving the truncated data.

With the mysql I am running on this causes the following exception on save.

taskflow.exceptions.StorageFailure: Failed updating flow details with uuid '193754c4-bf08-415c-9172-a54fec0bd690'
  DataError: (DataError) (1406, "Data too long for column 'meta' at row 1") 'UPDATE flowdetails....

[Ann Taraday (akamyshnikova)]

I guess this is fixed now as part of https://bugs.launchpad.net/taskflow/+bug/1838015