06-19-2017 10:52 AM
I've installed a Network Advisor 14.0.1 in SMI-Agent only for a customer to be able to interact with IBM Spectrum Control.
I've configured the SMI Agent to monitor multiple switch and I've activated the monitoring only for the ones I needed.
Everyhing is working pretty well except for the fact that in \install folder\bin I have multiple files named: hs_err_pidXXXX.mdmp that are creating themselves and filling up to 3gb each. At some point the drive is full and the monitoring stops.
Does anyone have any idea where it could come from ?
06-19-2017 11:51 PM
those MDMP are crash files from Java with error information. It might be that you have hit defect 597777 which is fixed in 14.0.2. Hard to say without having more information. Are you able to glean some information from the MDMP files? Did you change the max heap size, and if so to which value?
06-30-2017 08:08 AM
Im having this problem and Im on 14.0.1. The system is actually creating these .mdmp files almost everyday, and its in the "bin" directory of all places. What is the fix?
06-30-2017 11:45 AM
concerning defect 597777 - when the JVM crashes, it makes Network Advisor service to go down and come up frequently. The java crash hs_err file generated in the logs folder should have similar messages like below:
JVM crashed. A fatal error has been detected by the Java Runtime Environment: “EXCEPTION_ACCESS_VIOLATION"
So, please check hs_err file for error messages, if matches the above.
06-30-2017 02:15 PM
I looked in the log file, but these are being produced in the bin folder, but here is an excerpt of the top line
# A fatal error has been detected by the Java Runtime Environment:
# EXCEPTION_ACCESS_VIOLATION (0xc0000005) at pc=0x0000000071689eef, pid=2160, tid=2352
# JRE version: Java(TM) SE Runtime Environment (8.0_74-b02) (build 1.8.0_74-b02)
# Java VM: Java HotSpot(TM) 64-Bit Server VM (25.74-b02 mixed mode windows-amd64 compressed oops)
# Problematic frame:
# V [jvm.dll+0x4b9eef]
Now I have since increased MAXHEAP to 4096m, so just wanted to see if that helps. My only concern upgrading to 14.0.2 is the supportability with HPE Oneview 3.07
07-03-2017 02:00 AM
good. In the release notes, there is sizing requirement for the SMI/CIMOM part. If this is not working out, the opening a supportcase is the way to go, I belive since we would need much more data from the SMI Agent.
07-03-2017 12:58 PM
I upgraded to 14.0.3, and now I can discover my 2 switches???? ARGGGG It says authentication error, but I can SSH and HTTP into the switches. Why all of a sudden can I not discover my switches? Would turning on Target Driven Peer Zoning?