We have been facing issues on one of our PI interface machines (primary node), where-in after every weekly reboot, the network manager fails to start automatically. And based on our checks, it has been identified that the E drive which consists of PI home directory is n't ready/available when the network manager starts. From the event logs we have noticed that the PI network manager service starts first and fails, and after that the device (SAN disk) was ready and connected.
As the n/w manager service fails, the buffer ss and other PI interface services also fail to start.
We have checked this issue with the server team, and we have been suggested to add dependency of network manager service on the Microsoft iSCSI initiator service. This will ensure that the network manager service starts only when the iSCSI service starts, and in turn ensuring that the disk is available.
We have added this dependency as a workaround yesterday and we are planning to check during the next reboot, if this helps in proper restart of the service. The secondary interface node has no such issues, and is working as expected.
Meanwhile today, the pi-pi interfaces that were previously connected to secondary node, have failed over today to the primary node, and I could see a delay in data update from the interfaces. Am not sure if the delay is due to the change we 've made on the primary node, i.e. the dependancy of n/w manager service on iSCSI service.
Could someone please let me know if the issue with delay in data updates is related to the above change?
Also please let me know if the workaround suggested by server team is recommended by OSISOFT or not? If not, please let us know other possible solutions.