I've been doing some testing with the Fineract-CN,
using fineract-cn-docker-compose.

In previous testing of Fineract-CN before the postgres migration, I was
able to create and open loan products for members. However using the latest
versions of the microservices and provisioning using the scripts provided
in fineract-cn-docker-compose, I am not ables to open loan cases for
members.
I'm able to create members, create deposit and loan products, add these
products to members, and open the deposit account, but when attempting to
open the loan product, the portfolio-ms freezes up and doesn't respond to
any requests until it is restarted.

This is the relevant section from the portfolio-ms log:

[EL Fine]: sql: 2019-11-01
11:24:42.54--ServerSession(817490653)--Connection(1699212784)--Thread(Thread[qtp1761217448-17,5,main])--SELECT
t2.id, t2.a_comment, t2.executed_by, t2.executed_on, t2.case_id,
t2.task_def_id FROM bastet_cases t3, bastet_c_task_insts t2,
bastet_p_task_defs t1, bastet_products t0 WHERE (((t0.identifier = ?) AND
(t3.identifier = ?)) AND (((t1.id = t2.task_def_id) AND (t0.id =
t1.product_id)) AND (t3.id = t2.case_id)))
        bind => [test, test]
[EL Finest]: query: 2019-11-01
11:24:42.543--UnitOfWork(396538513)--Thread(Thread[qtp1761217448-20,5,main])--Execute
query ReadAllQuery(referenceClass=TaskDefinitionEntity sql="SELECT t1.id,
t1.actions, t1.description, t1.four_eyes, t1.identifier, t1.mandatory,
t1.a_name, t1.product_id FROM bastet_products t0, bastet_p_task_defs t1
WHERE ((t0.identifier = ?) AND (t0.id = t1.product_id))")
[EL Finest]: connection: 2019-11-01
11:24:42.544--ServerSession(817490653)--Connection(1400896017)--Thread(Thread[qtp1761217448-20,5,main])--Connection
acquired from connection pool [read].
[EL Finest]: connection: 2019-11-01
11:24:42.544--ServerSession(817490653)--Thread(Thread[qtp1761217448-20,5,main])--reconnecting
to external connection pool
[EL Finer]: transaction: 2019-11-01
11:24:42.548--UnitOfWork(441609303)--Thread(Thread[qtp1761217448-17,5,main])--release
unit of work
[EL Finer]: connection: 2019-11-01
11:24:42.548--ClientSession(782248230)--Thread(Thread[qtp1761217448-17,5,main])--client
released
[EL Finer]: connection: 2019-11-01
11:24:48.267--ServerSession(817490653)--Thread(Thread[qtp1761217448-17,5,main])--client
acquired: 1594588247
[EL Finer]: transaction: 2019-11-01
11:24:48.268--ClientSession(1594588247)--Thread(Thread[qtp1761217448-17,5,main])--acquire
unit of work: 1487145812
[EL Finest]: query: 2019-11-01
11:24:48.268--UnitOfWork(1487145812)--Thread(Thread[qtp1761217448-17,5,main])--Execute
query ReadAllQuery(referenceClass=ProductEntity )
[EL Finest]: connection: 2019-11-01
11:24:48.269--ServerSession(817490653)--Connection(1454951413)--Thread(Thread[qtp1761217448-17,5,main])--Connection
acquired from connection pool [read].
[EL Finest]: connection: 2019-11-01
11:24:48.269--ServerSession(817490653)--Thread(Thread[qtp1761217448-17,5,main])--reconnecting
to external connection pool
[EL Finer]: connection: 2019-11-01
11:24:48.28--ServerSession(817490653)--Thread(Thread[qtp1761217448-18,5,main])--client
acquired: 1954958561
[EL Finer]: transaction: 2019-11-01
11:24:48.281--ClientSession(1954958561)--Thread(Thread[qtp1761217448-18,5,main])--acquire
unit of work: 119694586
[EL Finest]: query: 2019-11-01
11:24:48.281--UnitOfWork(119694586)--Thread(Thread[qtp1761217448-18,5,main])--Execute
query ReadAllQuery(referenceClass=ProductEntity )
[EL Finest]: connection: 2019-11-01
11:24:48.282--ServerSession(817490653)--Connection(287564597)--Thread(Thread[qtp1761217448-18,5,main])--Connection
acquired from connection pool [read].
[EL Finest]: connection: 2019-11-01
11:24:48.282--ServerSession(817490653)--Thread(Thread[qtp1761217448-18,5,main])--reconnecting
to external connection pool

After "reconnecting to external connection pool", the microservice stops
responding. I assume this issue was introduced during the postgres
migration. Can anyone else recreate/resolve this issue?

Reply via email to