Update DBS specs to SQLAlchemy 1.1.13

[amaltaro]

This PR provides a new spec file for SQLAlchemy version 1.1.13, which hopefully will fix the issue seen in the DBS logs with the recent cx-Oracle upgrade to 7.3.0. Issue tracked here:
dmwm/DBS#645

This SQLAlchemy version is still expected to be using python byte strings (instead of unicode), so I expect it to solve the "twophase" issue and not to cause another one. But we can only be sure once it gets deployed.

DBS spec files have been updated to use this SQLAlchemy as well.

[muhammadimranfarooqi]

+1 to 4d63eda on slc7_amd64_gcc630.
Build log can be found here.

[amaltaro]

Hopefully a good sign, Imran completed the VM deployment this morning and I no longer see the sqlalchemy message that I saw in yesterday's DBS log:

/data/srv/beHG2102a/sw/slc7_amd64_gcc630/external/py2-sqlalchemy096/0.9.6/lib/python2.7/site-packages/sqlalchemy/engine/strategies.py:78: SAWarning: cx_Oracle is compiled under Python 2.xx using the WITH_UNICODE flag.  Consider recompiling cx_Oracle without this flag, which is in no way necessary for full support of Unicode. Otherwise, all string-holding bind parameters must be explicitly typed using SQLAlchemy's String type or one of its subtypes,or otherwise be passed as Python unicode.  Plain Python strings passed as bind parameters will be silently corrupted by cx_Oracle.
  dialect = dialect_cls(**dialect_args)

Let's see how it behaves once it gets pushed to testbed-k8s as well, in the coming few hours.

[muhammadimranfarooqi]

@amaltaro

I have deployed DBS. Seems to be working.

https://cmsweb-testbed.cern.ch/dbs/int/global/DBSReader/datasets?dataset=/ZMM*/*/*

Thanks

[amaltaro]

Perfect! Now we need to make sure there is no performance degradation.
@yuyiguo once it's convenient to you, please run the battery of dbs unit tests to have a better feeling of how DBS in cmsweb-testbed is performing.