Issues with Juniper devices on 0.8.7b?

[flailking]

Have been seeing an issue with only Juniper devices on 0.8.7b. When a change is made to a router, i.e., a new interface added or one turned down, the associated graphs are reporting NaN's and incorrect data. I have a server that is running 0.8.6i and it doesn't exhibit this behavior. New server is using Spine and the old server is using cmd.php. The action that corrects this behavior is to delete the host then re-add it, but only when all data sources are removed. As i had said this is only affecting juniper, all the cisco, extreme, alcatel, and others are fine. Attached is the output of the techsupport feature..

[rony]

Remove and readd the SNMP Interfaces Data Query on the host and set the "Re-Index Method" to "Index Count Changed".

[flailking]

I will wait for the next one to break and see if that fixes it. Should I apply this change to all the Juniper devices, or just make that change when it breaks? Also the older server is set for uptime/backwards...is this something that has changed between the versions?

thanks for the help

[rony]

That's a while back for me to recall what has happened in the code base to determine what has changed that might be causing your issue.