| T.R | Title | User | Personal
Name | Date | Lines |
|---|
| 961.1 | | a-109.tunnel.crl.dec.com::FORTY2::PALKA | Andrew Palka Altavista Directory | Mon Jan 20 1997 11:51 | 36 |
| 961.2 | Some feedback. | NETRIX::"[email protected]" | jan-ake olsson @ugo | Tue Jan 21 1997 08:20 | 52 |
| 961.3 | | a-101.tunnel.crl.dec.com::FORTY2::PALKA | Andrew Palka Altavista Directory | Tue Jan 21 1997 08:47 | 12 |
| 961.4 | Which files ? | NETRIX::"[email protected]" | jan-ake olsson | Tue Jan 21 1997 11:53 | 45 |
| 961.5 | | 16.120.224.4::FORTY2::PALKA | Andrew Palka Altavista Directory | Tue Jan 21 1997 11:57 | 8 |
| 961.6 | regarding the problem | ALEF::NIKOLIC | | Tue Feb 18 1997 20:58 | 13 |
|
I have a customer who says their DSA is also hanging (few details
available). I was wondering how logging was activated to see the
"2763:Exception reported at 477 in dsa_operators
Thread was alerted" message.
Also, if I get any info, I will send it to forty2::x500_support. If
there is a patch for the problem, I would welcome knowing about it in
this note string.
Thanks!
greg
|
| 961.7 | | a-101.tunnel.crl.dec.com::FORTY2::PALKA | Andrew Palka Altavista Directory | Wed Feb 19 1997 09:38 | 9 |
| 'Hanging' can mean several different things. Without knowing what
the symptoms are we can't say what kind of tracing would be
appropriate.
In general tracing should only be used under the direction of
engineering. It is not intended as a diagnostic tool for the
uninitiated.
Andrew
|
| 961.8 | y | ALEF::NIKOLIC | | Wed Feb 19 1997 14:56 | 15 |
|
I agree.
All I wanted to know was the note poster's trace settings to see if the
identical type of log entry is placed in my customer's log (i.e. if the
symptoms are the same).
I believe that my customer says he can get things to resume after
re-enabling the DSA but I suspect they are not very good at recording
events as they happen
Thanks and sorry for the confusion
greg
|
| 961.9 | | a-105.tunnel.crl.dec.com::FORTY2PALKA | Andrew Palka Altavista Directory | Thu Feb 20 1997 08:48 | 12 |
| The 'hang' from note .0 appears to be due to a search operation
taking a very very long time to complete, without being
interruptible. It is not possible to disable the dsa during this
time.
If your customer is able to disable an reenable the dsa then it
clearly is not the same hang.
That is why it is essential to find out exactly what is meant by
'hang'.
Andrew
|
| 961.10 | more info on the hang at my site | ALEF::NIKOLIC | | Mon Mar 10 1997 21:33 | 12 |
|
I actually logged into the system that was "hung" today. Don't know
the exact command that caused it to get that way. However, show
default on DXIM shows that DXIM is in the unbound state. Any command
to bind to the database (or do something like "show sub") never comes
back. I can FTP some files to my account. What might be useful to get
(i.e. files or outputs of some system commands?).
Thanks
greg
|
| 961.11 | | a-107.tunnel.crl.dec.com::FORTY2::PALKA | Andrew Palka Altavista Directory | Tue Mar 11 1997 08:39 | 19 |
| So, the symptoms are that you are unable to
bind to the dsa. Do you know if an existing association
with the dsa can continue to operate normally ?
This may be some problem with the Indaemon, which listens for
inbound connections. If this thread exits then you can't bind
to the dsa. The thread is restarted when you reenable the dsa.
There have been several problems in the past which have caused the
Indaemon to exit. The most recent one is due to some invalid
protocol being received during the bind process. This causes an
unexpected status code to be returned from OSAK, which was not being
handled.
Are there any events in the decnet event log ?
Is there anything in /var/dxd/dsa_output.log ?
Is this problem reproducible ? Do they have any idea what
triggers it ?
Andrew Palka
|