Re: [vwnc] [GS/SS Beta] Error using GsFile (Internal protocol error in the GCI to Gem connection)

Previous Topic Next Topic
 
classic Classic list List threaded Threaded
3 messages Options
Reply | Threaded
Open this post in threaded view
|

Re: [vwnc] [GS/SS Beta] Error using GsFile (Internal protocol error in the GCI to Gem connection)

Steve Aldred-3
Thanks for the suggestion Gerhard.  We tried topaz and it worked from
there.  We tried both VW7.5 and VW7.6 - both fail.

Further info as to what we're seeing when trying it from VisualWorks:

Writing/reading single character via #nextPut:/#next works:
    (GsFile openWrite: 'test.txt')
        nextPut: $a;
        close.

or

    (GsFile openRead: 'test.txt')
        next;
        close.

But doing both in the one execution results in the error (same error as
in initial email).

Writing or reading using #nextPutAll:/#next: results in the same error.

    (GsFile openWrite: 'test.txt')
        nextPutAll: 'a';
        close.

or

    (GsFile openRead: 'test.txt')
        next: 1;
        close.

It's almost as though there is some shared buffer that is being shared
between the regular file access and the GCI/Gem communication code.

Is anyone else using VW with GemStone and able to see this?

Thanks,
David and Steve

> When attempting to do file access using the following we get an
> exception ("GS Server Error - GbsLgcErrExpectedContinue - Internal
> protocol error in the GCI to Gem connection").  If we step over each
> line in the debugger it works fine.  In both cases we're using
> VisualWorks 7.6 to talk to GemStone.
>
> | f |
> f := GsFile openWrite: 'test.txt'.
> f nextPutAll: 'test64'.
> f close
>
> After the exception, the connection to the stone becomes unusable ("GS
> Server Error - GbsLgcErrPacketKind - Internal error: Gem to Client
> packet kind is unknown or unexpected") until we reconnect.
>
> Our configuration is as follows:
>
> 'processorCount'->4,
> 'gsRelease'->'2.2.5',
> 'osRelease'->'2.6.23.1-42.fc8',
> 'cpuKind'->'x86_64',
> 'gsBuildArchitecture'->'x86-64 (Linux)',
> 'gsVersion'->'2.2.5',
> 'osVersion'->'#1 SMP Tue Oct 30 13:18:33 EDT 2007',
> 'gsBuildDate'->'Fri Jan 25 15:16:13 2008',
> 'cpuArchitecture'->'x86-64'
>
> Anyone have any idea what the problem might be?

_______________________________________________
vwnc mailing list
[hidden email]
http://lists.cs.uiuc.edu/mailman/listinfo/vwnc
Reply | Threaded
Open this post in threaded view
|

Re: [vwnc] [GS/SS Beta] Error using GsFile (Internal protocol error in the GCI to Gem connection)

Eric Winger
Hi Steve,

Sorry for the delayed response. I don't see Gerhard's response included
in your mail, but I'm not seeing this problem. I tried running your
code snippet with vw76, windows clients against a GS/64 225 linux
server. I tried running the first two GsFile opens in one execution as
well as individually.

It looks like you sent the server configuration. Could you also let us
know what gci client libs you are using and what client OS's you are
using? And, can I assume the successful topaz test was run on the same
client OS & with the same gci libs?

thx

eric




On Apr 27, 2008, at 6:22 PM, Steve Aldred wrote:

> Thanks for the suggestion Gerhard.  We tried topaz and it worked from
> there.  We tried both VW7.5 and VW7.6 - both fail.
>
> Further info as to what we're seeing when trying it from VisualWorks:
>
> Writing/reading single character via #nextPut:/#next works:
>     (GsFile openWrite: 'test.txt')
>         nextPut: $a;
>         close.
>
> or
>
>     (GsFile openRead: 'test.txt')
>         next;
>         close.
>
> But doing both in the one execution results in the error (same error as
> in initial email).
>
> Writing or reading using #nextPutAll:/#next: results in the same error.
>
>     (GsFile openWrite: 'test.txt')
>         nextPutAll: 'a';
>         close.
>
> or
>
>     (GsFile openRead: 'test.txt')
>         next: 1;
>         close.
>
> It's almost as though there is some shared buffer that is being shared
> between the regular file access and the GCI/Gem communication code.
>
> Is anyone else using VW with GemStone and able to see this?
>
> Thanks,
> David and Steve
>
>> When attempting to do file access using the following we get an
>> exception ("GS Server Error - GbsLgcErrExpectedContinue - Internal
>> protocol error in the GCI to Gem connection").  If we step over each
>> line in the debugger it works fine.  In both cases we're using
>> VisualWorks 7.6 to talk to GemStone.
>>
>> | f |
>> f := GsFile openWrite: 'test.txt'.
>> f nextPutAll: 'test64'.
>> f close
>>
>> After the exception, the connection to the stone becomes unusable ("GS
>> Server Error - GbsLgcErrPacketKind - Internal error: Gem to Client
>> packet kind is unknown or unexpected") until we reconnect.
>>
>> Our configuration is as follows:
>>
>> 'processorCount'->4,
>> 'gsRelease'->'2.2.5',
>> 'osRelease'->'2.6.23.1-42.fc8',
>> 'cpuKind'->'x86_64',
>> 'gsBuildArchitecture'->'x86-64 (Linux)',
>> 'gsVersion'->'2.2.5',
>> 'osVersion'->'#1 SMP Tue Oct 30 13:18:33 EDT 2007',
>> 'gsBuildDate'->'Fri Jan 25 15:16:13 2008',
>> 'cpuArchitecture'->'x86-64'
>>
>> Anyone have any idea what the problem might be?
>
> _______________________________________________
> vwnc mailing list
> [hidden email]
> http://lists.cs.uiuc.edu/mailman/listinfo/vwnc
>

_______________________________________________
vwnc mailing list
[hidden email]
http://lists.cs.uiuc.edu/mailman/listinfo/vwnc
Reply | Threaded
Open this post in threaded view
|

Re: [vwnc] [GS/SS Beta] Error using GsFile (Internal protocol error in the GCI to Gem connection)

Steve Aldred-3


Eric Winger wrote:
> Hi Steve,
>
> Sorry for the delayed response. I don't see Gerhard's response
> included in your mail,

Gerhard suggested off list that we try it in Topaz, which worked.  We
were only able to run topaz on the server (which belongs to our client),
since we don't have any 64 bit machines.  Is there a 32 bit topaz that
will talk to a 64 bit server?

> but I'm not seeing this problem. I tried running your code snippet
> with vw76, windows clients against a GS/64 225 linux server. I tried
> running the first two GsFile opens in one execution as well as
> individually.
>
> It looks like you sent the server configuration. Could you also let us
> know what gci client libs you are using and what client OS's you are
> using? And, can I assume the successful topaz test was run on the same
> client OS & with the same gci libs?

Our client OS is Ubuntu 7.10 (32bit).  The client library we're using is
libgcirpc64-225.so.  Currently we haven't been able to test topaz and VW
on the same machine.  It may be possible to set up VW on our client's
server and try using that remotely to see if it has the same problem.

Our production system is going to be pure GemStone, so this is only a
development issue.  We can work around it by using topaz to run code
that does file access.  We're just puzzled as to why.

Thanks,
David and Steve

> On Apr 27, 2008, at 6:22 PM, Steve Aldred wrote:
>
>> Thanks for the suggestion Gerhard.  We tried topaz and it worked from
>> there.  We tried both VW7.5 and VW7.6 - both fail.
>>
>> Further info as to what we're seeing when trying it from VisualWorks:
>>
>> Writing/reading single character via #nextPut:/#next works:
>>     (GsFile openWrite: 'test.txt')
>>         nextPut: $a;
>>         close.
>>
>> or
>>
>>     (GsFile openRead: 'test.txt')
>>         next;
>>         close.
>>
>> But doing both in the one execution results in the error (same error as
>> in initial email).
>>
>> Writing or reading using #nextPutAll:/#next: results in the same error.
>>
>>     (GsFile openWrite: 'test.txt')
>>         nextPutAll: 'a';
>>         close.
>>
>> or
>>
>>     (GsFile openRead: 'test.txt')
>>         next: 1;
>>         close.
>>
>> It's almost as though there is some shared buffer that is being shared
>> between the regular file access and the GCI/Gem communication code.
>>
>> Is anyone else using VW with GemStone and able to see this?
>>
>> Thanks,
>> David and Steve
>>
>>> When attempting to do file access using the following we get an
>>> exception ("GS Server Error - GbsLgcErrExpectedContinue - Internal
>>> protocol error in the GCI to Gem connection").  If we step over each
>>> line in the debugger it works fine.  In both cases we're using
>>> VisualWorks 7.6 to talk to GemStone.
>>>
>>> | f |
>>> f := GsFile openWrite: 'test.txt'.
>>> f nextPutAll: 'test64'.
>>> f close
>>>
>>> After the exception, the connection to the stone becomes unusable ("GS
>>> Server Error - GbsLgcErrPacketKind - Internal error: Gem to Client
>>> packet kind is unknown or unexpected") until we reconnect.
>>>
>>> Our configuration is as follows:
>>>
>>> 'processorCount'->4,
>>> 'gsRelease'->'2.2.5',
>>> 'osRelease'->'2.6.23.1-42.fc8',
>>> 'cpuKind'->'x86_64',
>>> 'gsBuildArchitecture'->'x86-64 (Linux)',
>>> 'gsVersion'->'2.2.5',
>>> 'osVersion'->'#1 SMP Tue Oct 30 13:18:33 EDT 2007',
>>> 'gsBuildDate'->'Fri Jan 25 15:16:13 2008',
>>> 'cpuArchitecture'->'x86-64'
>>>
>>> Anyone have any idea what the problem might be?

_______________________________________________
vwnc mailing list
[hidden email]
http://lists.cs.uiuc.edu/mailman/listinfo/vwnc