[Sblim-issues] [ sblim-Bugs-1945381 ] NULL hostname causes SLP registration failure

[SourceForge.net <no...@so...>]

Bugs item #1945381, was opened at 2008-04-17 23:42
Message generated for change (Settings changed) made by mchasal
You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=712784&aid=1945381&group_id=128809

Please note that this message will contain a full copy of the comment thread,
including the initial issue submission, for this request,
not just the latest update.
Category: sfcb
Group: Function
>Status: Pending
>Resolution: Works For Me
Priority: 5
Private: No
Submitted By: Nicholas Bofferding (bofferdn)
Assigned to: Michael Chase-Salerno (mchasal)
Summary: NULL hostname causes SLP registration failure

Initial Comment:
[Problem]
SFCB is configured to fork during initialization and automatically register the WBEM SLP template properties with a local SLP service. If the hostname of the endpoint is set to "" (NULL) when the SFCB starts -or- when the SLP registration next refreshes, SFCB deregisters the SLP attribute string from the SLP service completely and never attempts to register it again, even if the user subsequently fixed the hostname.  It would be more friendly to try and reregister the service at each refresh interval (esp. for case where customer realized the problem and fixed the hostname).

$ hostname

$ sfcbd
--- sfcbd V1.2.3 started - 24030
--- Using /etc/sfcb/sfcb.cfg
--- Max provider procs: 10
--- initSocketPairs: 20
--- Max Https procs: 8
--- sfcbd HTTPS Daemon V1.2.3 started - 24032 - port 5989
--- Using Basic Authentication
--- Keep-alive timeout: 60 seconds
--- Maximum requests per connection: 10
--- localConnectServer started
--- Caching ClassProvider for /usr/var/lib/sfcb/registration/repository/root/ibmsd/classSchemas.gz (1.0-2) using 2800 bytes
--- Caching ClassProvider for /usr/var/lib/sfcb/registration/repository/root/interop/classSchemas.gz (1.0-2) using 812 bytes
--- HTTP-Daemon no longer active. Deregistering service with slp
--- Error deregistering service with slp (-22) ... it will now timeout

----------------------------------------------------------------------

Comment By: Michael Chase-Salerno (mchasal)
Date: 2008-06-05 18:29

Message:
Logged In: YES 
user_id=99742
Originator: NO

I've not been able to reproduce the crash of this, but here's what I've
found so far. A null hostname does cause the slp registration to be missing
some information, but fixing the hostname does eventually repair this. In
my case it took about 8 minutes after changing the hostname for the slp
data to be repaired. Here's my output:

[root@ ~]# hostname

[root@ ~]# /usr/local/sbin/sfcbd
--- sfcbd V1.2.3 started - 2023
--- Using /usr/local/etc/sfcb/sfcb.cfg
--- SSL not configured
--- User authentication disabled
--- Max provider procs: 32
--- initSocketPairs: 64
--- Max Http procs: 8
--- sfcbd HTTP  Daemon V1.2.3 started - 2025 - port 5988
--- Keep-alive timeout: 1 seconds
--- Maximum requests per connection: 10
--- localConnectServer started
--- Caching ClassProvider for
/usr/local/var/lib/sfcb/registration/repository/root/cimv2/classSchemas
(1.0-2) using 1120 bytes
--- Caching ClassProvider for
/usr/local/var/lib/sfcb/registration/repository/root/interop/classSchemas
(1.0-2) using 756 bytes

<NOTICE: no HTTP failure or deregistration from slp>

[msalerno@gubber ~]$ slptool unicastfindattrs someip  service:wbem
(template-type=wbem),(template-version=1.0),(template-description=This
template describes the attributes used for advertising WBEM
Servers.),(template-url-syntax=http://:5988),(service-hi-name=sfcb),(service-hi-description=Small
Footprint CIM Broker
1.2.3),(service-id=sfcb:NO-UUID-FILE-),(CommunicationMechanism=CIM-XML),(InteropSchemaNamespace=root/interop),(ProtocolVersion=),(FunctionalProfilesSupported=Basic
Read,Basic Write,Instance Manipulation,Association Traversal,Query
Execution,Indications),(MultipleOperationsSupported=false),(AuthenticationMechanismsSupported=Basic),(Namespace=root,root/cimv2,root/interop),(Classinfo=0,0,0)

<NOTICE: the URL and service-id are missing the hostname>

[msalerno@ SFCB_1_2_3]$ sudo hostname somehost.somedomain
[msalerno@ SFCB_1_2_3]$ hostname
somehost.somedomain
<and this output is somewhere between 6 and 8 minutes later:>

[msalerno@ ~]$ slptool unicastfindattrs someip  service:wbem
(template-type=wbem),(template-version=1.0),(template-description=This
template describes the attributes used for advertising WBEM
Servers.),(template-url-syntax=http://somehost.somedomain:5988),(service-hi-name=sfcb),(servic
e-hi-description=Small Footprint CIM Broker
1.2.3),(service-id=sfcb:NO-UUID-FILE-somehost.somedomain),(CommunicationMechanism=CIM-XML),(InteropSchemaNamespace=root/interop),(ProtocolVersion=),(FunctionalProfilesSupported=Basic
Read,Basic Write,Instance Manipulation,Association
Traversal,QueryExecution,Indications),(MultipleOperationsSupported=false),(AuthenticationMechanismsSupported=Basic),(Namespace=root,root/cimv2,root/interop),(Classinfo=0,0,0)

<NOTICE: also received this message in the sfcb output concurrent with the
slp update>
Callback Code -3

>From what I'm seeing this appears to be reasonable behavior. Perhaps
there's another facet that I'm not seeing that causes the crash. I did
manage to cause some problems myself when the mofc version didn't match and
cmpi-base wasn't rebuilt. If you have any other information that would help
here, I'll try it and see what happens on my end.

----------------------------------------------------------------------

You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=712784&aid=1945381&group_id=128809