Our devices are unable to cope with negative requests IDs, as a result, they go into timeout and a DataMiner restart is required to to reset the IDs. It is not feasible for to keep restarting DataMiner, hence we would like to request if it is not possible to prevent DataMiner from using negative numbers, can DataMiner not make improvements in that regard?
What other solutions are available to us if the above can not be realized? We would like to avoid timeouts due to negative requests IDs.
Hi Christine - to my knowledge this is not a lack of capabilities of DataMiner, but a bug in the SNMP stack of the device that needs to be managed, a bug that causes an incompatibility with the SNMP standard. This bug is somewhere in a third party SNMP stack that is used by various device vendors, and therefore it keeps on popping up once in a while. I believe that bug has also been fixed a long time ago, so maybe the devices have to be simply updated to their latest version?
Hi Thomas – I’ll double check but for what I recall our engineers cannot really work around this. DataMiner also uses a standard SNMP stack (i.e. we have not written our own version of the entire implementation of the standard SNMP protocol stack), and as this is really a bug causing a fundamental non-compliance with the SNMP standard, the failure already occurs in the low-level SNMP stack that we cannot change. But let me double-check, but if there would be an easy way to work around this, we would definitely consider it.
Hi, it looks like when using SNMPv3 in DataMiner, it should already limit the request ID ranging from 1000 to 32767. Not sure if SNMPv3 is an option in your case… SNMPv1 and v2 in DataMiner rely on a library where we cannot control the request ID. We might look at moving this to the library supporting the control of the request ID, but this is quite some work, so not sure if this would make it anytime soon.
We use the latest Kvarta Version & also requested Kvarta to look into this.
But we also notice this, that some vendors have this issue. So there is the question, if we can workaround this issue by excluding negative numbers from the request ids. So that a single devicegroup with this issue, doesn’t suddenly goes into “timeout” & require the whole dma to reboot.