There shouldn't be a difference between (in theory) how interfaces are enumerated by an snmpwalk and how open-nms does, however open-nms may be performing some specific snmpgets as a result of the output from the emuneration walk. Have you experimented with some random snmp gets, especially on invalid interfaces ?
--- Ray Van Dolson <[EMAIL PROTECTED]> wrote: > I'm running Net-SNMP 5.1.2-rc2 on a Red Hat 8.0 > system. This server functions > as a PPP Concentrator and thus has anywhere from > 400-500 PPP interfaces "up" > at any given time in addition to two Ethernet > interfaces. > > I am using OpenNMS 1.1.2 to monitor this server and > others like it. > > When OpenNMS runs its SNMP "discovery" process it > appears to query for all > the interfaces on the Concentrator system. Shortly > after the query begins, > Net-SNMP hangs and is non-responsive until the > daemon is killed and restarted. > > From observing the output with tcpdump, I see that > some responses get back > initially, but eventually cease altogether. I > started snmpd with the -D -f > -Le options to see if I could figure out what was > going on. Far too much > meaningless debug output for my eyes, but data was > certainly scrolling by at > a rapid rate even while the SNMP server appeared to > be hung as far as outside > programs were concerned. > > What scrolls by mostly is the following: > > trace: handle_getnext_loop(): snmp_agent.c, 2644: > results: trace: sprint_realloc_by_type(): > mib.c, 1971: > output: sprint_by_type, type 5 > . = NULL > > Literally hundreds of times at a very high rate. > Occasionally a large chunk > of "PPP PPP PPP" and so forth scrolls by and then > the previous block continues > looping in its place. > > I've brought this to the attention of the OpenNMS > people, but many of them > tell me they are monitoring boxes with that number > of interfaces with no > problems and suggesteted I check here for any > suggestions. > > A normal snmpwalk doesn't appear to freeze up > Net-SNMP. > > I should mention that I have tried both SNMPv1 and > SNMPv2c queries from > OpenNMS to see if the issue was exclusive to one > version or the other. No > difference however. > > Hope someone can give me some ideas! > > Ray Van Dolson > > > ------------------------------------------------------- > This SF.Net email is sponsored by OSTG. Have you > noticed the changes on > Linux.com, ITManagersJournal and NewsForge in the > past few weeks? Now, > one more big change to announce. We are now OSTG- > Open Source Technology > Group. Come see the changes on the new OSTG site. > www.ostg.com > _______________________________________________ > Net-snmp-users mailing list > [EMAIL PROTECTED] > Please see the following page to unsubscribe or > change other options: > https://lists.sourceforge.net/lists/listinfo/net-snmp-users > __________________________________ Do you Yahoo!? Yahoo! Mail - 50x more storage than other providers! http://promotions.yahoo.com/new_mail ------------------------------------------------------- This SF.Net email is sponsored by OSTG. Have you noticed the changes on Linux.com, ITManagersJournal and NewsForge in the past few weeks? Now, one more big change to announce. We are now OSTG- Open Source Technology Group. Come see the changes on the new OSTG site. www.ostg.com _______________________________________________ Net-snmp-users mailing list [EMAIL PROTECTED] Please see the following page to unsubscribe or change other options: https://lists.sourceforge.net/lists/listinfo/net-snmp-users
