System Monitor - socketRead0

[jkent]

On most of my transactions in the PTC System Monitor I always see socketRead0(FileDescriptor, byte[], int, int, int) as being the highest execution time.

Anyone have some insight in to what is causing the high execution time? Not sure where to start to track it down.

[Andrewdelk]

Hey man, I am a huge fan of PSM and I see that all the time. Here is the explanation I believe:

http://javaeesupportpatterns.blogspot.com/2011/08/javanetsocketinputstreamsocketread0.html

Hope it helps!

[jkent]

Thanks Andy. That looks to be the exact issue. Have you opened calls with PTC support on this? Did they come back with any resolution for you?

[Andrewdelk]

Gotta be honest with you, I sure haven't. I see it often and have been curious but I am sure you can relate when I say that I have time only to open tickets for the broken stuff right now in preparation for PTCPlanet.

This makes a lot of sense and my Dev team confirmed that it is legit. That's good enough for me

[jessh]

Looking at the caller breakdown it seems clear that the socketRead0's in question are JDBC calls waiting on responses from SQL Server. Thus the better question may be why your SQL Server is so slow to respond.

[jkent]

Here is an Add To Workspace that took 32.99 seconds. Of that 32.99 seconds 22.95 was spent in the method server and 10.04 was spent in SQL Server.