SM-FCFGM LB Comment Resolution, Rev 1

Keith Mccloghrie

Informations

• Publié par: Moveg
• Nombre de lectures: 11
• Langue: English

Extrait

SM-FCFGM Letter Ballot Comment Resolution, Rev 1
06-050v1, dated 2/16/2006
Company-# Tec Physical Section/table/figure Problem Description Suggested solution
hnic Page locator
al (not pdf
The subject of this ballot does not materially
Emulex-001 E general affect our organization.
Here's a big one. The year says 2005, but Change to 2006 in multiple places in the
McDATA-01 E 1 Copyright we're in 2006 now. document.
What does it meant for an Nx_Port to be
connected to Nx_Ports outside of the physical
McDATA-02 E 5 third paragraph fabric? Are you talking about NL_Ports? If so mention loops.
Shouldn't you mention platforms and their
McDATA-03 E 5 third paragraph properties too? Add a platform description.
Please mention that the Fabric Configuration
McDATA-04 E 5 Fourth paragraph Server is part of the Management Service. Add a new sentence.These groups are incorrect because a Fabric
Configuration Server should not be told to
discover. The Discovery is always real time
and replies should not be sent out unless the Delete or extremely modify these groups.
McDATA-05 T 5 5.4.2 and 5.4.3 local FCS is accessed. See more in McDATA-08/09.
The last paragraph shouldn't end on the next
McDATA-06 E 9 last sentence page. Connect the sentence.
At least correct spacing on From T11-TC-
MIB. It would probably be better to have the
IETF use longer lines, so have the IETF
McDATA-07 E 9 Imports Tabbing on Imports is odd. change the 72 character/line limit.t11FcsFabricDiscoveryStart doesn't seem to
correlate with GS. A Fabric Configuration
Server is only on a single Fabric. So you
shouldn't be able to tell it to discover on
multiple Fabrics. It seems odd to tell a Fabric
Configuration Server to start discovery since it
is part of Nx_Port login. Are you saying that
the IE will query end devices or are you saying
t11FcsFabricDiscove the management instance will query the
McDATA-08 T 16 ryStart Switches? Please clarify.
Delete this timer because it is implementation
t11FcsFabricDiscoveryTimeOut confuses me specific. The FCS must respond to the query
in a similar fashion to McDATA-08. There is no whenever it is made, but caching
time limit for how long the FCS stores requirements are not specified in T11. This
configuration data. The FCS should have it all means that the Discovery State Table should
of the time. The data could be cached by the be deleted as well. Usually, only local
management instance for a short amount of information is stored on each switch and
t11FcsFabricDiscove time, but the FCS stores it as long as it is distributed information is acquired at the time
McDATA-09 T 16 ryTimeOut online. of the query. This affects t11FcsIeTable also.
McDATA-10 T 19 t11FcsIeTable When or how is discovered data invalidated? Delete this reference.
is the Node WWN of the switch s/b is the
McDATA-11 E 20 t11FcsIeName Switch_Name of the switch
I think the formal Fibre Channel name should
be used when possible. This applies to
Fabric_Name, Switch_Name, Port_Name,
Logical Name, Platform Name etc. Fibre
Channel capitalizes names to give them
McDATA-12 T 20 t11FcsIeFabricName fabric name s/b Fabric_Name differentiation. It would be helpful.
information the ports s/b information on the
McDATA-13 E 25 t11FcsPortTable ports
McDATA-14 E 26 t11FcsPortName WWN of the port s/b Port_Name This distinguishes it from the Node_Name.
The reference to
t11FcsFabricDiscoveryTimeOut needs to be
McDATA-15 T 30 t11FcsPlatformTable removed.t11FcsTxRscns and t11FcsRxRscns should
not be counted by the FCS. This is more
related to the Name Server and sent from the
McDATA-16 T 38 t11FcsTxRscns Fabric Controller.
t11FcsReqRejectNot t11FcsReqectRejNotify s/b
McDATA-17 E 40 ifyEnable t11FcsReqRejectNotify
t11FcsDiscoveryCompNotifyEnable should
t11FcsDiscoveryCo never be sent because discovery is always real
McDATA-18 T 40 mpNotifyEnable time. Delete this notification.
A collection of objects for requesting a Fabric
t11FcsDiscoveryCon Configuration Server to discover the
McDATA-19 T 44 trolGroup configuration is not accurate. Remove this group.
t11FcsDiscoveryStat t11FcsDiscoveryStatusGroup doesn't make
McDATA-20 T 45 usGroup sense after all of the other comments.
Symantec-001 E 5 3 In the 2nd sentence of the 5th paragraph, "The Fabric Configuration Server is designed
delete the second "it can be". so that it can be distributed among switches
and accessed from any Nx_Port.
Symantec-002 E 5 3 In the 2nd sentence of the last paragraph, "The latest version is [FC-GS-5]."
replace "specification" by "version".
Symantec-003 E 7 5.3 The sentence "With multiple fabrics, each Clarify.
fabric has its own mangament
instrumentation." needs clarification. Separate
agents?
Symantec-004 E 7 5.3 The first sentenced of the last paragraph is Suggest "In a fabric conformant to FC switch
unclear. standards prior to FC-SW-4……."Symantec-005 T 14 6 The description of the Clarify
t11FcsFabricDiscoveryTable is the first
instance of the use of the term "locally-
managed switch". This needs some
clarification, and possible some definition in
section 5. I also don't remember this term
being used before. Is it intended to mean the
switch on which the FCS is implemented etc.
Symantec-006 T 16 6 The description of the Clarify
t11FcsDiscoveryStateTable uses the term
"locally-managed switch".
Symantec-007 T 18 6 In the last paragraph of the description of the Clarify and add additional text if necessary.
t11FcsDiscoveryStatus object, is the discovery
aborted ONLY if it was initiated from this
switch?
Symantec-008 T 19 6 In the Description of the t11FcsIeTable, please Clarify
clarify what is meant by "remote topology". The
term is not apparently used in GS-5.
Symantec-009 T 23 6 The description of the t11FcsMgmtListEntry Clarify
uses the term "locally-managed switch".
Symantec-010 T 24 6 The description of the t11FcsPortListEntry uses Clarify
the term "locally-managed switch".
Symantec-011 T 25 6 The description of the t11FcsPortEntry uses Clarify
Symantec-012 T 29 6 The description of the Clarify
t11FcsAttachPortNameListEntry uses the term
"locally-managed switch".
Symantec-013 T 30 6 The description of the t11FcsPlatformTable Clarify
uses the term "locally-managed switch".
Symantec-014 T 30 6 The description of the t11FcsPlatformEntry ClarifySymantec-015 T 34 6 The format in GS-5 of the Management Change reference to include the Management
address referenced in the Address Value in GS-5 section 6.2.3.2.7
t11FcsPlatformSysMgmtAddr object includes a
length field? Is this length field included in the
object?
Symantec-016 T 34 6 The format in GS-5 of the Management
address referenced in the Address Value in GS-5 section 6.2.3.2.7
t11FcsPlatformCLusterMgmtAddr object
includes a length field? Is this length field
included in the object?
Symantec-017 T 35 6 The description of the Clarify
t11FcsNodeNameListEntry uses the term
"locally-managed switch".
Symantec-018 T 36 6 The description of the t11FcsStatsEntry uses Clarify
the term "locally-managed switch".
Symantec-019 E 20 8 Reference Scott as T11.5 Vice Chair T11.5 Vice-Chair: Scott Kipp, McData. T11.5
members.
Symantec-020 T 21 9 FS-NS-MIB missing from list of references. Add referenceeditor's response agreement at T11.5 status as of
meeting on 2/8/2006 2/16/2006
accepted no change
accepted agreed done
accepted -- it's not
referring to NL_Ports; will
rephrase to change "ports
outside the fabric" to
"Nx_Ports" agreed done
accepted -- will add a
sentence: "Each platform
may have attributes
including its name …" agreed done
accepted agreed donerejected -- these groups
are an optional means to
manage "local data copies"
as described in SW-4;
specifically, GS-5's section
B.6.4.2 (informative) points
to SW-3, and section 9.4.1
of SW-4 says "The dMS
within the Entry Switch
services the request by
making any needed
requests of other dMS
contained by the other
Switches, if the required
information is not available
locally" and "A dMS may
maintain local data copies,
and a dMS shall notify
other dMS that they should
remove local data copies".
Also note that FCS
commands (section 6.2.5
in GS-5) have lots of Gets
but only a few Registers so to be reviewed by the
info can not be distributed submitter of the
at registration time, comment pending
accepted agreed done
accepted agreed doneto be reviewed by the
submitter of the
related to McDATA-05 comment pending
to be reviewed by the
submitter of the
related to McDATA-05 comment pending
submitter of the comment pending
accepted agreed done
accepted agreed done
accepted agreed done done
to be reviewed by the
submitter of the
related to McDATA-05 comment pendingagreed to remove
RSCN-related
counters from this
MIB (and add to