| T.R | Title | User | Personal
Name | Date | Lines |
|---|
| 3833.1 | Not that I understand this Wave stuff | AIMTEC::WICKS_A | Atlanta's Most (In)famous Welshman | Mon Jan 31 1994 15:46 | 7 |
| Peter,
Didn't something similar get discussed in 3767?
Regards,
Andrew.D.Wicks
|
| 3833.2 | | ZUR01::KURTH | Peter Kurth @RLE, R�mlang (Switzerland) | Wed Feb 02 1994 10:23 | 13 |
| Hello
No, it isn't the same as note 3767. It seems, that the FCS cannot
understand the cluster alias and displays %OAFC-E-INVAUTHENT.
Other sites have also DECnet/OSI installed, the object 73 is set
with "Outgoing Alias = True" and FCS works with no problems.
Something must be set wrong here. Any ideas?
Currently, I workaround this problem by setting the "Outgoing Alias" to
False during execution of A1V30_SITE_START.COM.
|
| 3833.3 | | ZUR01::KURTH | Peter Kurth @RLE, R�mlang (Switzerland) | Wed Feb 02 1994 14:44 | 33 |
| Hi
Some more information:
When doing NCL>SHOW SESS CONT PORT * ALL WITH Remote End User
Address = number = 73, the 'links' between the user and the FCS looks
like this:
With Outgoing Alias=False
Client = <Default value>
Local End User Address = UIC = [0,0]ALLIN1
Transport Port = OSI Transport Port OSI$PORT_007E
Direction = Outgoing
Remote End User Address = number = 73
Node Name Sent = EMD:.GRD.BERN.GRDBA2
With Outgoing Alias=True
Client = <Default value>
Local End User Address = UIC = [0,0]ALLIN1
Transport Port = OSI Transport Port OSI$PORT_01E1
Direction = Outgoing
Remote End User Address = number = 73
Node Name Sent = EMD:.grd.bern.grdbe1
As seen, the 'Node Name Sent' is different (GRDBA2 is the node, grdbe1
is the alias).
Is the FCS not able to handle the small letters from the cluster alias
grdbe1? Does the validation happens case sensitive?
Any hint?
Thanks and regards ,Peter
|
| 3833.4 | Could be the lowecase alias... | IOSG::STANDAGE | | Wed Feb 02 1994 14:59 | 12 |
|
Peter,
I'm not sure how the FCS reacts to lowercase cluster alias's, but I know
from personal experience that other products do not stand up well to
this (Message router I believe). Redefine the alias to be in uppercase,
stop/start the servers and see what happens...
Kevin.
|
| 3833.5 | | ZUR01::KURTH | Peter Kurth @RLE, R�mlang (Switzerland) | Wed Feb 02 1994 15:37 | 24 |
| Hi
Re -1: Redefining the alias is not so easy. We would have to change the
DNS name space; and this may be difficult in a large network (as my
DECnet/OSI colleague said). And case sensitivity seems not to be the
cause, see below.
More info:
On good old Phase IV, do the following:
NCP> set obj obj_73 alias outgoing enable
go into ALL-IN-1 and do MGT MFC MS and you get "%OAFC-E-INVAUTHENT,
Invalid authentication information received by the File Cabinet Server".
Huhh, what is this?
So, this is a problem with the cluster alias in general, not with
Phase V. Phase V is worse, because the object is set Outgoing Alias=True
per default. But, it seems, that other sites have no problem with FCS
and DECnet/OSI.
More clues?
Regards, Peter
|
| 3833.6 | | CSOA1::LENNIG | Dave (N8JCX), MIG, @CYO | Wed Feb 02 1994 21:19 | 10 |
| re: .0
Is FCS knowledgable re Phase V nodenames?
Perchance this
Node Synonym = False
should be this?
Node Synonym = True
Dave
|
| 3833.7 | Not a phase V problem, a cluster alias problem | ZUR01::KURTH | Peter Kurth @RLE, R�mlang (Switzerland) | Thu Feb 03 1994 14:24 | 10 |
| Hi Dave
I played with Node Synonym. Setting it to False or True does not
make any difference.
The question remains: Why can FCS not work with cluster alias outgoing
enabled (it can be DECnet Phase IV or V)? Can somebody have a look at
a DECnet Phase V/ALL-IN-1 system, what are the settings?
Regards, Peter
|
| 3833.8 | Fixed in V3.1 ... | ZUR01::KURTH | Peter Kurth @RLE, R�mlang (Switzerland) | Tue Jul 12 1994 16:17 | 6 |
| Final status:
This is fixed in ALL-IN-1 V3.1 (ICO#1950). For V3.0, use the workaround
from note 3767.9.
Regards, Peter
|