Question

WADACO not failing over even though IFS does with Oracle RAC?

  • 26 October 2020
  • 6 replies
  • 240 views
NickPorter
Rank
Userlevel 6
Badge +18

We’re seeing an issue when Oracle fails over or moved from one RAC node to another.  IFS itself works as expected but Warehouse Data Collection (WADACO) does not. 

The result is that anyone attempting to use WADACO in either in the IFS IEE screens or in the ScanIT touch apps receives an error telling them that "no Warehouse Data Collection licenses are available" but this is not correct.  Checking the WADACO licenses we see that we have many free each time this has happened.

We can fix by restarting the IFS Apps Server services using mws-svr command but this causes a dull downtime for all IFS that we need to avoid.

Has anyone got any insight into why WADACO isn’t failing over like IFS does?

With the mechanism in place such that WADACO only consumes a license when a user accesses those screens inside IFS makes me think that the WADACO connections are made differently than IFS as a whole.

6 replies

william.klotz
Rank
Userlevel 7
Badge +21

Hi @NickPorter ,

 

I will be interested in seeing the answer to your question.  We will be implementing a RAC solution in the near future and we use WADACO as well.

 

One area you might want to look into is our mobile application installation uses a separate Oracle database for the mobile application configurations and uses a an IIS installation as the reverse proxy.  You may wish to look into if you have a separate Oracle database installation for your mobile applications.

 

Regards,

William Klotz

D
Rank
Userlevel 6
Badge +16

Hi Nick,

I don't know anything about RAC, but I don't think wadaco is handling connections much differently than the rest of the application.

I think the reason you get license error here is because of the custom licensing functionality that wadaco is using to check the wadaco user licenses that is very sensitive to a stable and working connection to the IFS App Server and if there is any issues there, we unfortunately catch issues like that in the licensing handling that are not really license issues and instead environment issues.

We have seen these environment issues causing license errors in our internal dev environments from time to time, but I have not heard about any real big issues with customer environments so far.

This should not be a problem in the coming IFS Cloud 21R1 release since licensing is handled differently there.

Mohamed Infaz
Rank
Userlevel 5
Badge +10

Hi Nick,

One of our customers recently had this issue, are your both RAC nodes are in sync on time? The customer had a mismatch of an hour and correcting it solved the issue. The reason you are getting licenses are full is that there are a lot of Active/Parked sessions sitting idle because of the larger issue you are facing. When these sessions reach the maximum license number it throws you the error. Aren’t your end-users getting randomly locked by another user error messages too?

Cheers, Infaz.
NickPorter
Rank
Userlevel 6
Badge +18

@infaz 

No, the issue for us is not the time sync.  Both server nodes are identical.

The cause of the “no licenses” message in this case is NOT because of parked or idle sessions.  As noted that was  the first thing that was looked at before realizing there was a failover issue.

No, no other locks or errors.

I created a ticket with IFS when this came up and we realized the root cause.  After review I was advised that IFS R&D would look at making improvements to the failover process, which to me means they identified a gap.  In the meantime they developed a patch that doesn’t fix the failover issue, but it does instead provide a better error message.

Nick

Mohamed Infaz
Rank
Userlevel 5
Badge +10

Hi Nick,

Thanks for letting me know, how frequent is this error?

Cheers, Infaz.
NickPorter
Rank
Userlevel 6
Badge +18

@infaz This only happens when we failover a RAC node, whether intentionally or through a failure, so not often.  Now that we know how it functions we just have to plan to do a restart of the Apps Server services to refresh the connection for WADACO.

Sadly this means that although the failover works to keep IFS up and running, we then have to do a controlled restart of IFS soon after to fix WADACO because that is a critical item for parts of our business… 

Reply


Terms & ConditionsCookie settings