We have production issue for posting document to TN.
We have few clustered wM server. They are assessing same TN database.
Once in a while (may be once in a week), our document posting Application doesn’t get response back (i.e. code 200) from TN receive service. It brings down performance of all the servers (same day).
First we doubted on the TN database.
But after many days’ research done by DBA, everything looks good so far in Database.
Now we don’t have any clue - Why all Server’s performance reduces at the same day? The only common thing is Database. But So far didn’t found any issue.
RMG, THANKS for quick reply.
Yesterday I was looking at Threads and memory usage. It was OK. They were not at peak. But normally we have very high traffic during Monday compare to other days.
since cluster and accessing the same DB…any database locks and is indexing in place?? My first thought was also the DB…You already check the memory/threads usage…
What is the current Logging level set to in Production??
RMG,
Just now I talked to Prod. support people. Sometimes thread count reaches the peak (450). By default watt.server.cache.flushMins=10. Do you think reducing it to “5” will help ?
Thanks.
It’s not adviceable to reduce by default…But whats the setting you have for JAVA Min/Max heap set to in the server.bat?? may be its worth to look at it and update accordingly (max it to 1536) depends on your OS/jvm esp Prod.
When the issue happens, you may check the Service Usage page, see which service is running for long time. this may give you clue about where it is holding up. take snap shot of the page every few minutes, then compare them. Most likely it’s still DB related.
Tong