[dhi-servers] NameServer communication problems

Philip Crotwell crotwell at seis.sc.edu
Mon Aug 20 09:03:28 PDT 2007 

Hi

You should not need to run multiple copies of the servers. We  
regularly copy all the references from your name server to one we run  
here. So all you would need to do would be to set up the omninames,  
then copy the server references from the existing name server to the  
new one.

Philip

http://www.seis.sc.edu/fisheye/browse/seis/trunk/fissuresUtil/src/edu/ 
sc/seis/fissuresUtil/namingService/NameServiceCopy.java?r=17652

On Aug 20, 2007, at 10:32 AM, Sue Schoch wrote:

> Morning again Philip,
>
> Never mind my last question.  I've been out of dhi for a while.  Of  
> course I can run both nameservers at once, but I'd have to run  
> multiple copies of the servers to connect to the omni nameserver.
>
> Sue
>
> On Aug 19, 2007, at 9:21 PM, Philip Crotwell wrote:
>
>>
>> Hi
>>
>> Interesting that that was it. I know figuring it out for our  
>> clients was a time consuming process. The underlying issue is that  
>> the orbacus orb is old enough that when it was written there  
>> really weren't any character sets used except ascii. Now, many  
>> os's use utf8, and the client orbs respect whatever settings the  
>> client has set. This is of course a nice thing if you are a non- 
>> english speaker and want to send text via corba. But the use of  
>> utf8 by newer orbs apparently has the downside off not working  
>> with the older orbacus orb.
>>
>> Unless the server-side orb is upgraded, the only solution I know  
>> of is for the clients to explicitly set their character set to ascii.
>>
>> Sorry,
>> Philip
>>
>> On Aug 18, 2007, at 5:30 PM, Charles Ammon wrote:
>>
>>> Hi,
>>>
>>> I contacted the author of ADORB (Victor Ananiev) and he seems to  
>>> have
>>> figured out what is going on. The issue is the character code set
>>> translation (good call Philip).
>>>
>>> If Victor is right, it appears that the server restart was not  
>>> exactly
>>> what was running before Wednesday. Please see the response from him
>>> below. He makes several suggestions for me to work around the issue,
>>> but if possible, the easiest for the group of people running my
>>> clients would be to make the adjustment at the server to be the same
>>> as it was before restart on Wednesday.
>>>
>>> Would this be possible? If not, I can start on the workarounds.
>>>
>>> Chuck
>>>
>>> ---------- Forwarded message ----------
>>> From: Victor Ananiev <v_ananiev at mac.com>
>>> Date: Aug 18, 2007 12:51 PM
>>> Subject: Re: ADORB stopped talking to another orb
>>> To: Charles Ammon <cammon at geosc.psu.edu>
>>>
>>>
>>> Hi Charles,
>>>
>>> The problem is in the character codeset negotiation. For some reason
>>> they don't allow UTF8 (CORBA code 0x0501001).
>>> It connects if I send ASCII (0x00010020) or no codesets at all.
>>> Apparently they have changed something in the configuration or,
>>> according to their log, forgot to register a converter for UTF8.
>>>
>>> For the time being you can change the value for native char code in
>>> the file IDLTaggedComponent.m
>>> //static int _nativeCharCode = 0x05010001;
>>> static int _nativeCharCode = 0x00010020;
>>>
>>> I don't know how this can affect the rest of your application  
>>> though.
>>>
>>> I will look for a better solution meanwhile.
>>> The other way is to remove codesets from the service context using
>>> interceptors that I have added in the version 1.5.
>>>
>>> Regards,
>>> Victor Ananiev
>>>
>>>
>>> On Aug 18, 2007, at 9:04 AM, Charles Ammon wrote:
>>>
>>> Hi,
>>>
>>> I am sorry to bother you. I am an ADORB user and you helped me  
>>> figure
>>> some things out about using CORBA a long time ago. Since then, I  
>>> have
>>> been developing and using ADORB to talk with a seismic data server -
>>> and all has been running well until this week. Now I can't  
>>> connect to
>>> the nameserver I have been talking to for more than a year. They did
>>> restart the machine running the server, but they don't think that
>>> anything has changed - it must have. I have codes that I have not
>>> recompiled that used to work, but don't any more...
>>>
>>> I downloaded the most recent version of IDLBrowser and have been
>>> trying to connect to the server, but all I get is a time out. I  
>>> appear
>>> to be the only one who is having problems (all others use java-based
>>> orbs to communicate).
>>>
>>> The server address is 128.95.166.129:6371. Using IDLBrowser, I have
>>> tried talking to the server with GIOP 1.0, 1.1, and 1.2 (1.2 used to
>>> work). All time out.
>>>
>>> I have been in touch with the server administrators and this is the
>>> error that they see when I try to connect:
>>>
>>>   id: iiop
>>>   local address: 192.168.167.2:6371
>>>   remote address: 128.118.174.210:49185 ]
>>> com.ooc.OB.AssertionFailed: ORBacus encountered an internal error
>>>         at com.ooc.OB.Assert._OB_assert(Assert.java:19)
>>>         at com.ooc.OB.CodeSetDatabase.getConverter 
>>> (CodeSetDatabase.java:230)
>>>         at com.ooc.OB.GIOPServerWorker.getCodeConverters 
>>> (GIOPServerWorker.java:1
>>> 00)
>>>         at com.ooc.OB.GIOPServerWorker.executeRequest 
>>> (GIOPServerWorker.java:441)
>>>         at com.ooc.OB.GIOPServerWorker.execute 
>>> (GIOPServerWorker.java:314)
>>>         at com.ooc.OB.GIOPServerWorkerThreaded.receiverRun 
>>> (GIOPServerWorkerThrea
>>> ded.java:521)
>>>         at com.ooc.OB.GIOPServerWorkerThreaded$ReceiverThread.run 
>>> (GIOPServerWork
>>> erThreaded.java:70)
>>> com.ooc.OB.AssertionFailed: ORBacus encountered an internal error
>>>         at com.ooc.OB.Assert._OB_assert(Assert.java:19)
>>>         at com.ooc.OB.GIOPServerWorkerThreaded$ReceiverThread.run 
>>> (GIOPServerWork
>>> erThreaded.java:75)
>>>
>>> If you have any ideas, can you please let me know?
>>>
>>> Thank you,
>>> Chuck Ammon
>>> ==================================================================== 
>>> ==
>>> Charles J. Ammon, Department of Geosciences
>>> Penn State University / 440 Deike Bldg / University Park, PA 16802
>>> VOICE: (814) 865 2310  / FAXES: (814) 863 7823 or (814) 863 8724
>>> http://eqseis.geosc.psu.edu/~cammon/
>>> http://www.personal.psu.edu/faculty/c/j/cja12/
>>>
>>>
>>>
>>>
>>> -- 
>>> ==================================================================== 
>>> ==
>>> Charles J. Ammon, Department of Geosciences
>>> Penn State University / 440 Deike Bldg / University Park, PA 16802
>>> VOICE: (814) 865 2310  / FAXES: (814) 863 7823 or (814) 863 8724
>>> http://eqseis.geosc.psu.edu/~cammon/
>>> http://www.personal.psu.edu/faculty/c/j/cja12/
>>> _______________________________________________
>>> dhi-servers mailing list
>>> dhi-servers at iris.washington.edu
>>> http://www.iris.washington.edu/mailman/listinfo/dhi-servers
>>
>> _______________________________________________
>> dhi-servers mailing list
>> dhi-servers at iris.washington.edu
>> http://www.iris.washington.edu/mailman/listinfo/dhi-servers
>
> _______________________________________________
> dhi-servers mailing list
> dhi-servers at iris.washington.edu
> http://www.iris.washington.edu/mailman/listinfo/dhi-servers

More information about the dhi-servers mailing list