Data Source database connection drop

David Fennell shared this problem 2 years ago
Resolved

One of our clients has an issue where Yellowfin's Data Source connection to our SQL database will 'drop' or 'fail' despite none of the login credentials changing. This shows on the Admin Console page as a failure to connect to the datasource - see attached screenshot


To resolve, we simply access the data source page and test the connection. It connects successfully and now the reports are able to load.


We have experienced this on YF 7.1 20160325 and also after upgrading to YF7.1 20161209. This is a new issue that has only starting occurring in the past month, whereas the client has been using Yellowfin since July 2015. Yellowfin and the SQL are installed on the same server.


Have you seen this issue before? Do you have any suggestions why this might be occurring?

Comments (8)

photo
1

Hi David,

I hope all is well,

So I can investigate further can you please read through the below information/requests:

First of all can I question if you have looked at this (old) forum post explaining timeout settings etc; http://portal.yellowfin.bi/YFForum-Yellowfin-DB-Configuration-settings-Timeout-and-Max-Open-Connections-?thread=106876

This should prove to be a good read and may help this issue.

With this can I also question the following:

-Do you have any long running queries which might be "killing" the MySQL server connection?

-Did this only start occurring after some system change or new report/broadcast?

-Can you please send across all logs files from the \yellowfin\appserver\logs folder and let us know roughly when the connection has dropped out (to help us hunt for a specific period) .


Despite the above I can see that you are running a version of which is no longer supported; 7.1 can I question if you plan on upgrading at soe point to our latest 7.2 or 7.3? I will await your reply.


Regards,

Mark

photo
1

Hi Mark

It's a Microsoft SQL database, not MySQL, if that makes any difference at all.


There are some newer long running reports. How would these kill the SSMS connection?


I'll see if I can find out when the connection dropped so that I can send you the logs.


You say that 7.1 is no longer supported, but a new version was released on 8th Feb for 7.1, 7.2 and 7.3 ?

photo
1

Hi Mark. Log files have been sent to support@yellowfin.bi under ticket 254214

photo
1

Hi David,


Thanks for getting in touch with me regarding this, I can confirm I have received the Logs, I will begin the process of looking through these logs. I do apologise for the delay.


Regards,

Mark

photo
photo
1

Hi David,


I hope all is well and sorry for the delay in responding to you. I have since looked into this further and discussed this with my fellow support colleagues. From what we can see we would like to look towards Volatile data sources as we believe this may help with your issue. If you can please check the below links out I would appreciate this greatly:


https://portal.yellowfin.bi/YFForum-Volatile-sources-mechanism-and-repository-DB-Connection-?thread=147681

http://portal.yellowfin.bi/YFForum-Determining-closed-connections-?thread=140652


I look forward to your reply.


Regards,

Mark

photo
1

Hi David,


I hope all is well,


I just want touch base and question whether or not you are in need of any further assistance with regards to this issue. If you can please get back to me asap I would appreciate this greatly. I look forward to your reply.


Regards,

Mark

photo
1

Hi Mark

The client reports that it has been resolved, and turned out to be an internal issue:

“I looked through the logs myself before my holidays and I think I may have solved the problem. There was a problem with authentication. Around 3.30 am the following error was logged:”java.sql.SQLException: Login failed. The login is from an untrusted domain and cannot be used with Windows authentication.”, after which the datasource was marked as unavailable. At 3.30 am our primary domain controller was restarted a couple of times a week. I don’t know why the secondary domain controller didn’t take over but when I checked, it hadn’t been restarted since shortly before the problems started. To keep on the safe side, I changed the times the servers are restarted at night so that the domain controller is restarted before the Union Square server. I asked Emma whether the reports dropped off during my absence and she said they didn’t.

So when I was checking whether the reports were available in the morning, I must have skipped a few days, even though I thought I had checked them each day. So in my mind the reports were sometimes available in the morning and then they dropped off later in the day, which turns out not to be true as it transpired from the logs. Sorry for the confusion this created.”

photo
1

Hi David,

Thank you for getting back to me with this good news. I am pleased all is well and working how it should be. If you have any further problems please do not hesitate to submit a new problem/ticket within our community and we will gladly assist where we can. I will now go ahead and mark this problem as resolved.


Regards,

Mark