Question

Schedules more query to fme_node_health and fme_lock table


Badge +1

Hi all

I have FME Flow installed in production environment with 2 nodes (Fault-Tolerant installation).

I have now 4 workspaces that i have scheduled every 4 minutes.

My DB Administrator see many query like this 

UPDATE fme_node_health SET modifieddate = :1 , expirydate = :2 , timeout = :3 , expiry = :4  WHERE process = :5  AND name = :6

UPDATE fme_lock SET val = :1 , uuid = :2 , expiry = :3  WHERE lock_key = :4  AND (uuid = :5  OR expiry < :6 )

They occours more time and and fill the disk space dedicated to Oracle logs.

It's the first time it's happened to me and I don't understand if it depends on the scheduling or the Fault-Tolerant type installation.

Can someone help me?

Thanks a lot


5 replies

Userlevel 5
Badge +29

What frequency are you seeing these updates? Looking at a sever I have access to I can see it updating various rows every ~15s.

Seems appropriate as these table appear to be keeping track of the various components of the Fault Tolerant setup

Badge +1

Thank you very much for the reply.

In fact the frequency is similar to yours.

The problem of the Oracle logs then disappeared: in reality we were releasing many components into production and when checking some things the DB admin raised this question to me because he saw a lot of writings by FME.

Another question: is it possible to configure this write parameter to the DB that you know of?

Thanks

Userlevel 5
Badge +29

Having a cursory look through the config files, I can’t find anything that seems to control that frequency.

I really don’t think it should change, if anything I’m surprised it only appears to checking fault tolerance every ~15s, I’d have expected it to be more frequent than that.

@rylanatsafe or @siennaatsafe might have some input or be able to get someone else from Safe to chime in

Badge +9

Hi @nicol.ricci and @hkingsbury,

 

Thanks for tagging us Hamish. 

@nicol.ricci I spoke to our support team about this and it would be helpful if you could create a case. You can create a case for us here: https://support.safe.com/hc/en-us/p/Support

There’s likely some improvements we can make in our internal health check process to not log to the database, or to do it less frequently. 

Badge +1

Hi 

I will talk about this with my team and probably i will open a case.

Thank you all

Reply