JVM Clock Drift

We are using webMethods 8.2.2. We have noticed that the time (for example: getCurrentTime) is drifting. We have two servers with a webMethods IS cluster. The servers have identical times. After we recycle the ISes, the times match. After its been up a week, its off by 15 seconds.

I found this entry for the JVM to use the system clock instead of the stock Java method based upon number of CPU systems since restart.

-XX:+UseGetTimeOfDay

I put this in the setenv.sh file in the JAVA_ARG1 setting.

Has anyone used this setting? Where do I put it in to make it work? I guessed setenv.sh.

Also, a quick search revealed that using this setting could degrade performance. Anyone experience this?

We are using HP/UX.

Thanks!
Brian

Please check these 2 articles from the Empower and decide your options:

https://empower.softwareag.com/sl24sec/SecuredServices/KCFullTextASP/viewing/view.asp?KEY=097338-15926433&DSN=PIVOTAL&DST=TCD&HL=1&QUERY=XX:+UseGetTimeOfDay&SessionID=973459866

https://empower.softwareag.com/sl24sec/SecuredServices/KCFullTextASP/viewing/view.asp?KEY=102094-9087058&DSN=PIVOTAL&DST=TCD&HL=1&QUERY=XX:+UseGetTimeOfDay&SessionID=973459866

I put this in the setenv.sh file in the JAVA_ARG1 setting.

Has anyone used this setting? Where do I put it in to make it work? I guessed setenv.sh.

–> Yes that is where the place the java arguments go:

HTH,
RMG

Thank you, RMG!

Glad to know I placed it in the right file! We have it in QA now. Prod soon…

NP…So I assume the setting resolved the time diff?

HTH,
RMG

After 24 hours in QA, the clocks between the two clustered instances still match. If they match in a week, I will be really happy!

OK good sign: :smiley: