Good afternoon friends.
We have an amd with memory leak problems, always in the morning from Monday to Friday.
When it happens we have to restart the rtmhs process, with the loss of traffic that this entails.
Have you ever had this problem?
Thank you very much.
Solved! Go to Solution.
We have hard time to answer this question with only this bite of data.
Anyway, NAM/DC RUM 2017 is out of support now, for far more than a year. Our advice is to upgrade to 2018 or 2019 ASAP and verify whether the issue still exists.
Please note that NAM product support generally ends March 31, 2021 - in just five weeks. This means that you need to hurry. Open a Support ticket now, download supported version of the NAM, request current license, and upgrade!
Also, since version 2017 is not supported anymore, the upgrade may turn out into a fresh installation. But that may make all things simpler, as the time is running out.
Thank you very much for your answer. We know we are out of support.
We just wanted to share this problem in case it had happened to someone and find a solution
We are currently restarting the rtmhs process when the memory leack occurs, although we have an estimated loss of 30 min.
I still very highly recommend upgrading to the current 2019 sp5, as this issue does not seem to exist in the latest release.
I believe I wrote a script that would monitor the creation of the zdata files, and if the NAM Probe was more than a couple minutes late producing the current zdata file, kill the process and let the systemd watchdog restart it. Staying on the leaking version is guaranteed to have data loss when the leak trigger is present; this script only reduces the data loss to an interval rather than the 30-60+ minutes of lost data you have waiting for the leaking/hung process to crash.