Hi, we are running on AX 2012 R3. The Server configuration is the following: - 3 AOSs in load balancing running for rich client connections - 2 AOSs Non load balanced , for the batch queues execution All servers have 24 GB of RAM and 8 virtual processors. We have currently 200 average sessions (User + Worker) for each load balanced AOS. We are not doing live customization import, we are using Model-store importation scheduled once every two week, do all the checks before starting the system and so the system is always stable when it's running live. On Sunday morning 02:00 CET , we have a scheduled DB resync that did not ended in 24 hours so my colleague killed the spid of the running job. Yesterday happened something very unusual in our environment: Apparently randomly, the load balanced AOSs started to crash and we had a chain on crashes that i think happened because of servers saturation. Load balancer started to not work well and plus on that we started to get this error message for all the users "Failed to create a session; confirm that the user has the proper privileges". After some checks, the only way I could go back on a stable environment was to : 1. Turn off all the aoss istances 2. Turn on one aos istance and do a full CIL 3. Run the DB-Sync The DB-Sync, on the middle check, threw 1488 warnings like this: Index: 'xxxx' may cayse unique index violations. What may have caused this issue, and what we can do to prevent it to happen again? Any suggestions is very appreciated, Kind regards,
↧