image crash
Locked
 
|
Hello! I'm experiancing a crash the moment I start the image. The stacktrace makes it clear that it fails in a primitive while trying to close a stale dbx connection. The database is not even running, let alone having a valid connection to it.
I get this quite often. Can you think of a solution for it, please? $ ./pharo.sh DBXFatalError: FATAL OpenDBX: Invalid handle DBXPostgresBackend(DBXBackend)>>errorCode:handle:
DBXPostgresBackend(DBXBackend)>>verifyError:handle: DBXPostgresBackend(DBXBackend)>>closeConnection: DBXConnection>>close DBXConnection>>disconnect DBXConnection>>finalize
WeakFinalizerItem>>finalizeValues WeakRegistry>>finalizeValues in Block: [finalizer finalizeValues] BlockClosure>>on:do: BlockClosure>>on:fork: in Block: [Processor terminateActive]
Milan Mimica http://sparklet.sf.net |
|
|
On Wed, Jun 20, 2012 at 11:03 PM, Milan Mimica <[hidden email]> wrote:
Do you know what is the value of the handle at this point?
-- Mariano http://marianopeck.wordpress.com |
|
|
On 20 June 2012 23:08, Mariano Martinez Peck <[hidden email]> wrote:
Let's say I could find out, if I empower my gdb skills. I will try. I expect it to be a value of a handle that was valid at some point in time... What do you hope it is?
Milan Mimica http://sparklet.sf.net |
|
|
In reply to this post by Mariano Martinez Peck
Which image/vm are you using? I remember experiencing vms where the conections were not gced :/
On Wed, Jun 20, 2012 at 11:08 PM, Mariano Martinez Peck <[hidden email]> wrote:
|
|
|
It's a Pharo 1.4 one click from jenkins, from 2 weeks ago or so.
On 20 June 2012 23:15, Guillermo Polito <[hidden email]> wrote: Which image/vm are you using? I remember experiencing vms where the conections were not gced :/ Milan Mimica http://sparklet.sf.net |
|
|
In reply to this post by mmimica
On Wed, Jun 20, 2012 at 11:14 PM, Milan Mimica <[hidden email]> wrote:
mmmm right....that should be from gdb :(
I don't know, maybe nil ? as a workaound you can just disable the automatic disconnection of connections (done in the shutDown).... does this crash also happens when disconnecting connections in the normal scenario (not with the GC) ? -- Mariano http://marianopeck.wordpress.com |
|
|
In which OS are you? have you compiled OpenDBX yourself or use a binary distribution?
const char* odbx_error( odbx_t* handle, int error ) { DEBUGLOG( if( handle != NULL ) { handle->log.write( &(handle->log), 1, "odbx_error() called" ); } ) if( error == -ODBX_ERR_BACKEND ) { if( handle != NULL && handle->ops != NULL && handle->ops->basic != NULL && handle->ops->basic->error != NULL ) { return handle->ops->basic->error( handle ); } return dgettext( "opendbx1", odbx_errmsg[ODBX_ERR_HANDLE] ); } if( error <= ODBX_ERR_SUCCESS && error >= -ODBX_MAX_ERRNO ) { return dgettext( "opendbx1", odbx_errmsg[-error] ); } return dgettext( "opendbx1", odbx_errmsg[ODBX_ERR_PARAM] ); } On Wed, Jun 20, 2012 at 11:19 PM, Mariano Martinez Peck <[hidden email]> wrote:
-- Mariano http://marianopeck.wordpress.com |
|
|
On 20 June 2012 23:24, Mariano Martinez Peck <[hidden email]> wrote: In which OS are you? have you compiled OpenDBX yourself or use a binary distribution? Using slackware linux. I have OpenDBX compiled myself. The version probably is the latest 1.4.5, but the shared objects have the 1.2.0 suffix!?
I can't tell why the image quits. It doesn't even crash, it exists with code 01. And what are you trying to say with this code? When is odbx_error even called?
Milan Mimica http://sparklet.sf.net |
|
|
In reply to this post by Mariano Martinez Peck
On 20 June 2012 23:08, Mariano Martinez Peck <[hidden email]> wrote:
It's 135887552 Milan Mimica http://sparklet.sf.net |
|
|
In reply to this post by Mariano Martinez Peck
On 20 June 2012 23:24, Mariano Martinez Peck <[hidden email]> wrote: In which OS are you? have you compiled OpenDBX yourself or use a binary distribution? Ah, I think you missunderstood where the image quits. It doesn't crash, don't mind the subject.
"odbx_error" is called just fine. DBXBackend>>rrorCode:handle: signals a DBXFatalError which makes the image quit. So the error is handled gracefully on the dbx side. It is Pharo's start-up process that is still weak.
Milan Mimica http://sparklet.sf.net |
«
Return to DBXTalk
|
1 view|%1 views
| Free forum by Nabble | Edit this page |