Re: IPS WG Last Call: iSNS MIB
Keith McCloghrie <[email protected]>
| Newsgroups | gmane.ietf.ips |
|---|---|
| Message-ID | <[email protected]> |
> This message announces an IP Storage (ips) Working Group > Last Call on the iSNS MIB draft: > > Definitions of Managed Objects for iSNS > (Internet Storage Name Service) > > draft-ietf-ips-isns-mib-08.txt > > This WG Last Call will end at 12:00 midnight (end of day) Eastern > Time on Friday, March 17, 2006. All technical comments and > requests for changes need to be posted to this list. Editorial > comments and corrections may be sent directly to the draft > editor, Scott Kipp <[email protected]> and cc:'d to me > (David Black <[email protected]>) as WG chair. See below. Keith. ----------------- 1. Incorrect usage of MODULE-COMPLIANCE: > isnsServerComplianceV1 MODULE-COMPLIANCE > STATUS current > DESCRIPTION > "This set of groups is required for full implementation > by an iSNS Server if it has the resources to keep > track of the number of registered objects in iSNS Server > instances over time." > MODULE -- this module > MANDATORY-GROUPS { > isnsServerNumObjGroup > } > ::= { isnsCompliances 3 } First, MODULE-COMPLIANCE statements are not accummulative. In other words, each MODULE-COMPLIANCE, including the one quoted above, specifies the minimum requirements needed to claim a particular level of compliance. The one above says it is "for full implementation by an iSNS Server", and it requires support for just a single object group. Therefore, any iSNS Server which supports the isnsServerNumObjGroup can claim to be a full implementation of the MIB, irrespective of any other groups/objects it might/might not support. This is clearly wrong. Second, a MODULE-COMPLIANCE is the wrong place to put an "if" clause. To resolve both of these, change the above MODULE-COMPLIANCE into a GROUP clause, which specifies isnsServerNumObjGroup as Conditional Mandatory, and include this GROUP clause in (presumably) each of the remaining MODULE-COMPLIANCE clauses: isnsIscsiServerComplianceV1 & isnsIfcpServerComplianceV1. 2. Compilations errors/warnings: - the position of isnsSrvrInstCntrlNodeAuth in its SEQUENCE doesn't match its OID; move it to come after isnsSrvrInstEnblCntrlNdeMgtScn in the SEQUENCE, - you are required to have at least one read-able object in every table. In your MIB, isnsCntlNodeFcPortTable & isnsRegIscsiNodeTable only have one object which is not-accessible. You must change the one object in both cases to be read-only, and because they are now read-only, they now need to be included in an OBJECT-GROUP. - MODULE-COMPLIANCE appears twice in the IMPORTS clause; delete the duplicate. - delete the items in the IMPORTS which are never used: Gauge32, ZeroBasedCounter32 ZeroBasedCounter64, InterfaceIndexOrZero, PhysicalIndexOrZero & StorageType 3. The Abstract section: - if I recall correctly, the Abstract section is not supposed to include references. 4. References - section 2 begins: The iSNS protocol can be used by IP based storage devices for dynamic registration and discovery of storage devices in the network [RFC 471]. I suspect this reference should be to [RFC4171] ?? - when the unused TC's (see above) are removed, then a number of the References will no longer be referenced and need to be removed. 5. IANA Considerations - delete "4371" from: ::= { transmission 4371 } it is inappropriate to list unassigned values. - I don't think this qualifies as a MIB which should go under the "transmission" subtree -- the rule of thumb for the transmission subtree is that the "{ transmission nnn }" OID is assigned to the media-specific MIB (see section 4 of RFC 2863) for the type of interface to which ifType=nnn has been assigned. I suggest that this MIB ask for an assignment under the "mib-2" subtree. To do this, just change each occurrence of "transmission" to "mib-2". 6. Editorial: - isnsRegIscsiNodeEntry's DESCRIPTION includes the words: ... The RowStatus managed object provides a method to delete registered nodes ... but there is no RowStatus object in this MIB. Delete the sentence. - I personally don't like the indentation and spacing scheme used in the MIB, but since the same scheme is used very consistently, it's not a problem. - I personally don't have a problem with the OID hierarchy in this MIB, but Appendix D of RFC 4181 recommends a different layout. _______________________________________________ Ips mailing list [email protected] https://www1.ietf.org/mailman/listinfo/ips