Ebenezer, Please see replies inline.
On Sun, Mar 11, 2018 at 4:27 AM, Ebenezer Graham <[email protected]> wrote: > Hello there, > > I want to setup Fineract CN. I checked the updated setup guide and step > three requires me to install MySQL 5.7. > > Now, this is my worry: I have Mifos X on my PC and installed MySQL 5.5 as a > result of the complications I had with 5.7. Is it possible to use 5.5 > instead of 5.7 for Fineract CN > I can't comment on this specifically so I'll let others do so. However for Mifos X, an upcoming release, Fineract 1.2 (which powers Mifos X) will provide support for MySQL 5.7. > > Please advise me on what to do. > > Note: I can't afford to run two SQL servers at the moment. Also, can I work > on Fineract CN without experimenting with Mifos X? I can get rid of the > Mifos X however, I would like to have it and continue studying it whiles I > working on Fineract CN. Can I keep both in the system and use the same > MySQL server (5.5)? > You could work on Fineract CN without experimenting with Mifos X but if you're looking to continued exploring both, I'd first wait on whether or not you can run Fineract CN on MySQL 5.5. If not, I'd suggest running MySQL 5.7, exploring Fineract CN and then waiting for the updated Fineract 1.2 release which will be compatible with MySQL 5.7. We'll also soon have a public demo server of Fineract CN up too which will help with exploration of that. > > *Best Regards,* > > *Ebenezer Graham* > > *BSc (Hons) Computing* > > > [image: EmailSignature.png] > > Beau Plan Business Park | West Building > > Pamplemousses | Mauritius > > > <http://www.pactmart.com/> > <http://www.pactmart.com/> > www.pactmart.com | Freelancing made easy. > > skype: > ebenezer.graham > | Phone: > +230 5840 9206 > GitHub <https://github.com/ebenezergraham> | LinkedIn > <https://www.linkedin.com/in/ebenezer-graham/> | Twitter > <https://twitter.com/pactmart> | Facebook > <https://www.facebook.com/pactmart> | Personal Website / CV > <http://ebenezergraham.github.io/> > > > *“Practice isn’t the thing you do once you’re good. It’s the thing you do > that makes you good.”* > -*Malcolm Gladwell* > > > > On 6 March 2018 at 03:54, Viswa Ramamoorthy < > [email protected]> wrote: > > > Hello Myrle, > > Thanks for sharing your thoughts on Dockerization of services. > > My experience with embedding infrastructure inside application JVM (even > > for development purposes) has not been great. In a development, when > things > > change so much, embedding infrastructure adds additional time to > bootstrap > > the whole thing when application restarts needed. Having external > > infrastructures gives better visibility as well as their failure to start > > can be diagnosed better (e.g. a port is not available because another > > instance of a infrastructure is already running in the background). > > With external infrastructures, installation becomes cumbersome if we go > > with installation of infrastructure and every one need to follow those > > steps to install to get there. My PR is really to solve that part. > > Some of the complexity, that you alluded to, are really complexity of > > design/developing in micro services architecture. > > Couple of points about logging (that stays within Docker) as well as > debug > > mode with Docker deployment, are very much solvable with Docker > deployment. > > Regarding high amount of resources needed for deployment, one strategy > > that could be looked into is to provide capability to selectively start > > services needed for a feature to complete and leave the full deployment > to > > integration environments. > > If you looking into collapsing micro-services into a single war, for > > development purposes, it can be a strategy that would work. But all of > the > > services need to be using compatible version of frameworks and managing > > different configurations can be challenge. > > Having infrastructure as Docker can still come handy in day to day > > development. I understand the timeline/priority. No problem. > > RegardsViswa > > > > On Monday, March 5, 2018 07:17:41 AM EST, Myrle > > Krantz <[email protected]> wrote: > > > > Hey Viswa, > > > > It's going to take me a little longer to get to merging and reviewing > > this, so please be patient with me. But a couple of comments while > > you're waiting: > > > > 1.) That you're not seeing those error messages probably may not mean > > they are gone. It may mean that they are now "hidden" in the docker > > image. That's not ideal for error messages. It makes debugging > > harder when there really is an issue. > > 2.) Thank you for finding the error with the artifact path. Consider > > submitting a patch to fineract-cn-service-starter. > > > > I'm a bit concerned about the idea of moving this all into docker. > > Yes docker is one important method for deploying microservices, and > > showing an example of how to use those technologies is important. But > > the demo-server is also there partly to test code and get a local > > installation up and running. When I started on it, my intention was > > to support Mark van Veen so that he didn't have to start all the > > services and then provision by hand to work on the UI. Unfortunately > > there are serious problems with the demo-server the way it is now. It > > takes a huge amount of resources because it starts every service in > > its own process. Many developers do not have computers with > > sufficient resources to run this locally. At one point, Kuelap > > literally bought me a new computer after I had spent a couple of days > > unsuccessfully trying to make the demo-server work because Markus had > > added a couple more services to it. Moving these processes into > > containers doesn't solve this problem. Docker works with computing > > resources in a shockingly efficient manner, so it probably doesn't > > actually make the problem worse, but it does make the problem harder > > to solve. > > > > Another point, is that currently I can start these services in debug > > mode, and attack a debugger to them to understand tricky problems. I > > don't know how to do that in a docker container. Any changes in that > > direction should consider this use case. > > > > I can see that testing this running in docker might be important for > > some of our users and contributors. But I don't want it to be the > > default. I would feel more comfortable with your change sets if you > > made this "more optional". > > > > My first priority for this project is to enable contributors. To do > > that, I'd like to look for ways to run all of the services in one > > process for the purposes of local testing and debugging. > > > > Best Regards, > > Myrle Krantz > > Committer, Apache Fineract > > > > > > > > On Fri, Mar 2, 2018 at 3:35 AM, Viswa Ramamoorthy > > <[email protected]> wrote: > > > Hi, > > > I have raised a PR with docker compose yml for Eureka and ActiveMQ. > > > It is https://github.com/apache/fineract-cn-demo-server/pull/3 > > > Please note that after I launch Eureka and ActiveMQ via Docker, I do > > not see JMS connect error as well as Eureka registration error anymore. > > > But service launch was failing with below errorCould not find artifact > > io.mifos.provisioner:service-boot:jar:0.1.0-BUILD-SNAPSHOT > > > Locally was able to fix artifact path to "org.apache.fineract.cn." in > > fineract-cn-service-starter and move forward with service launch. > > > But there were more errors. I have not looked into further yet. > > > I think demo server needs some more work to get it to work > consistently. > > All of the services can be launched via shell script if there are no > start > > up dependencies between them > > > Regards > > > Viswa > > > > > -- *Ed Cable* President/CEO, Mifos Initiative [email protected] | Skype: edcable | Mobile: +1.484.477.8649 *Collectively Creating a World of 3 Billion Maries | *http://mifos.org <http://facebook.com/mifos> <http://www.twitter.com/mifos>
