"we didn't pay enough attention" is a very kind way to put it. The entire SNMP exercise was about monitoring; those like me who tried to make things writable pushed rocks uphill politically. That, and the experience of trying to make event triggered management rather than poll-driven management, is the major reason that I, personally, don't want to have anything to do with network management in the IETF.
On Apr 21, 2010, at 4:53 AM, Bill Fenner wrote: > [sorry if you get duplicates, replying from an address that actually > *works* this time] > > 2010/4/19 Tina TSOU <t...@huawei.com>: >> In RFC 4293, >> - ipAddressTable is described as writable, this table uses address >> as index, but the critical information for configuring address, the >> address prefix ipAddressPrefix node is read-only. It seems >> contradict to me. > > Worse, the whole ipAddressPrefixTable is read-only, so you can't > insert a new row to try to point to. This seems like we didn't pay > enough attention to making these configurable via SNMP. > >> This table uses address as index, but the public network and VPN >> may have exactly the same IP address. This table does not >> explicitly say whether it only supports public network. > > As with most MIBs, if you have multiple contexts on the device, you > need multiple instances of the MIB, usually using SNMPv3 contexts and > the entLogicalTable to associate SNMPv2 communities if needed. > > Bill > -------------------------------------------------------------------- > IETF IPv6 working group mailing list > ipv6@ietf.org > Administrative Requests: https://www.ietf.org/mailman/listinfo/ipv6 > -------------------------------------------------------------------- http://www.ipinc.net/IPv4.GIF -------------------------------------------------------------------- IETF IPv6 working group mailing list ipv6@ietf.org Administrative Requests: https://www.ietf.org/mailman/listinfo/ipv6 --------------------------------------------------------------------
