[MXS-1422] JIRA doesn't work through MaxScale Created: 2017-09-18  Updated: 2017-09-21  Resolved: 2017-09-21

Status: Closed
Project: MariaDB MaxScale
Component/s: N/A
Affects Version/s: 2.1.7
Fix Version/s: N/A

Type: Bug Priority: Major
Reporter: Igor V. Lovalev Assignee: Johan Wikman
Resolution: Not a Bug Votes: 0
Labels: None
Environment:

Centos 6 x86_64



 Description   

Hi all! Not a bug, i just want to ask you for your experience: Are you install JIRA with maxscale? I have maxscale 2.1.7 cluster with 3 percona-mysql nodes, and i try to install JIRA with it, but i failed... if i try install JIRA without maxscale (just one master percona-mysql) - all ok. I can't understand where is trouble. So i think maybe you had same troubles...



 Comments   
Comment by markus makela [ 2017-09-18 ]

Please attach your MaxScale configuration file with all sensitive information removed. With it, I believe we'll be able to rule out which part might be causing the problems. If Jira produced any errors, please attach those as well.

To my knowledge, there should be no reason why MaxScale would not work with Jira. In addition, I would classify this as a bug since it works without MaxScale but fails with MaxScale.

Comment by Igor V. Lovalev [ 2017-09-18 ]

Hi again, Markus!
Thanx 4 ur answer! I don't think this maxscale bug, i just ask for it here, cuz your team and your project is awesome, and i think i find help here (... and cuz you too use jira ofc)
Anyway, i'll make test stand with jira and maxscale and attach all debug info here soon!

Comment by Johan Wikman [ 2017-09-20 ]

Hi iglov
I know nothing about the Jira installation process, so I am just speculating here...

If the Jira installation process involves writes that are immediately followed by reads of that same data and the assumption is that what was written will be found, then, if the written data has not been replicated by the time it is read, I can imagine that the Jira installation process gets confused.

If you want to experiment you could check if the installation succeeds with the CCR filter installed.

Comment by Igor V. Lovalev [ 2017-09-20 ]

Hi, Johan & Markus!
I think i resolve the trouble I have 2 clusters: very fast production cluster (1 maxscale, 1 master percona server + 2 slaves) and very slow testing cluster (1 maxscale, 1 master percona server + 4 slaves). All this time, i tried to install jira on fast cluster where database creation time is ~30sec and finally it's failed, but today, for more debugging, i tried install jira on slow cluster where database creation time is ~10-20min and... it's finished without errors. Then, i read comment by Johan and had seen new feature for me (I didn't know about this before) - CCR filter; and i thought: - Hmm, it's may work's for me! After that i took a risk - i apply CCR on production and tried install jira again. And what do you think? It's worked! Now jira installs his database on fast cluster without errors. It's amazing!
And i want to say you guys, big thank for your attention and time, you awesome

Comment by Johan Wikman [ 2017-09-21 ]

Hi Igor, good to hear that CCR did the trick for the production cluster! I'll close this as Not a Bug then

Generated at Thu Feb 08 04:06:37 UTC 2024 using Jira 8.20.16#820016-sha1:9d11dbea5f4be3d4cc21f03a88dd11d8c8687422.