I wondered that as well, but I tried to simulate the not found and error code-paths, but I couldn't trigger it. So I'm not ruling it out, I just can't reproduce it.
Another thing that's weird is that it looks like the index has been stripped from sensorStatus, which might be an indication that weird is going on inside libagentx. But like I said: without a reproducer I haven't been able to pin it down. So the additional verbose information should be useful. Come to think of it: The `sysctl hw.sensors` output might be helpful as well, both on a succeeding run, as well as at the time of the crash (maybe something like: `while true; do date; sysctl hw.sensors; sleep 1; done > \ /path/to/output`) @Ryan: this info doesn't need to be on the list, so feel free to send it to me in private if you want. On Fri, 2022-10-28 at 11:08 +0100, Stuart Henderson wrote: > I wonder if there are any sensors which disappear and reappear.. > > On 2022/10/28 10:01, Martijn van Duren wrote: > > Could you run snmpd with `-vv`? That way I also have the specific > > OIDs being requested and returned (both frontend and backend) and > > might make it a little more easy to reproduce. > > > > Do note that this adds at least 4 log lines for every request > > issues to snmpd, so your logfile might explode a bit. > > > > martijn@ > > > > On Thu, 2022-10-27 at 14:08 -0700, Ryan Freeman wrote: > > > On Thu, Oct 27, 2022 at 01:46:21PM -0700, Ryan Freeman wrote: > > > > Hello, > > > > After upgrading some virtual machines to OpenBSD 7.2, I started noticing > > > > snmpd dying approx every 6 hours on the upgraded machines. > > > > > > > > Oct 27 13:14:33 mirror snmpd[98795]: AgentX(1268939451/2580462718): > > > > 2506302838 > > > > iso.org.dod.internet.private.enterprises.openBSD.sensorsMIBObjects.sensors.sensorTable.sensorEntry.sensorStatus: > > > > oids not equal > > > > Oct 27 13:14:33 mirror snmpd[98795]: AgentX(1268939451/2580462718): > > > > Closing: Too many parse errors > > > > Oct 27 13:14:33 mirror snmpd[98795]: AgentX(1268939451/2580462718): > > > > Closed by snmpd (Too many AgentX parse errors from peer) > > > > Oct 27 13:14:33 mirror snmpd_metrics[88325]: [fd:0 sess:2580462718 > > > > ctx:<default>]: unsupported call: agentx-Close-PDU > > > > Oct 27 13:14:33 mirror snmpd[98795]: AgentX(1268939451): Connection > > > > reset by peer > > > > Oct 27 13:14:33 mirror snmpd[98795]: snmpe: AgentX(1268939451): > > > > disappeared unexpected > > > > > > > > The message is always the same, it tends to be around 1:20am, 7:20am, > > > > 1:20pm, 7:20pm > > > > I have a script set to check "rcctl ls failed" and notify if something > > > > has failed. > > > > > > > > LibreNMS is used to scrape the snmpd instances on the affected VMs. > > > > > > And, forgot to include the snmpd.conf, apologies. here it is with minor > > > changes values only: > > > # $OpenBSD: snmpd.conf,v 1.2 2021/08/08 13:43:10 sthen Exp $ > > > > > > # See snmpd.conf(5) for more options (tcp, alternative ports, trap > > > listener) > > > listen on 127.0.0.1 > > > > > > user "changed" auth hmac-sha1 authkey "randomstuff" enc aes enckey > > > "morerandomstuff" > > > > > > # Adjust the local system information > > > system contact "Systems Team ([email protected])" > > > #system location "Rack A1-24, Room 13" > > > > > > # Required by some management software > > > system services 74 > > > > > > LibreNMS then scrapes it using snmpv3 and authPriv mode. > > > no core file is being dropped by snmpd > > > > > > -Ryan > > > > >
