Community Discussions and Support
mercury/32 crash

Hi

 this happen two/three times per day and always same address 0X0001008b

V.
<P>Hi</P> <P> this happen two/three times per day and always same address 0X0001008b</P><FONT face=Tahoma size=1><FONT face=Tahoma size=1>V.</FONT></FONT>

Hi

sometimes (not too many) the mercury/32 MTA crash with the following info window:

 Runtime 216 at 00080194 

and it wait for a click to restart..

how solve it?? if not possible, how to skip automatically the click waiting to avoid the automatic restart??

 

<P>Hi</P> <P>sometimes (not too many) the mercury/32 MTA crash with the following info window:</P> <P> Runtime 216 at 00080194 </P> <P>and it wait for a click to restart..</P> <P>how solve it?? if not possible, how to skip automatically the click waiting to avoid the automatic restart??</P> <P mce_keep="true"> </P>

What version of Mercury? 

Is the address changing?

What else do you run on this system?

<P>What version of Mercury? </P> <P>Is the address changing?</P> <P>What else do you run on this system?</P>

the configurarion is:

windows 2003 sp2

mercury/32 4.62

stunnel

this is a production system with a fixed ip (internal and external) and no changes since mercury upgrade from 4.5x to 4.62 in Jul 2008 

 

<P>the configurarion is:</P> <P>windows 2003 sp2</P> <P>mercury/32 4.62</P> <P>stunnel</P> <P>this is a production system with a fixed ip (internal and external) and no changes since mercury upgrade from 4.5x to 4.62 in Jul 2008 </P> <P mce_keep="true"> </P>

The runtime error is definitely not a Mercury/32 error and I suspect the stunnel.

The runtime error is definitely not a Mercury/32 error and I suspect the stunnel.

[quote user="vmgracia"]

sometimes (not too many) the mercury/32 MTA crash with the following info window:

 Runtime 216 at 00080194 

[/quote]

Just to confirm Thomas's reply - this is not a Mercury error message, nor is it generated by the Mercury runtime library (the Borland libraries don't put up runtime exception notifications).

It *might* be SpamHalter, but only Lukas would know that - I have no way of helping you with that. It might also be something else in the system, such as STunnel (as mentioned by someone else).

Sorry I can't give more insight, but I can tell you this isn't Mercury crashing per se.

Cheers!

-- David --

[quote user="vmgracia"]<p>sometimes (not too many) the mercury/32 MTA crash with the following info window:</p> <p> Runtime 216 at 00080194 </p>[/quote] Just to confirm Thomas's reply - this is not a Mercury error message, nor is it generated by the Mercury runtime library (the Borland libraries don't put up runtime exception notifications). It *might* be SpamHalter, but only Lukas would know that - I have no way of helping you with that. It might also be something else in the system, such as STunnel (as mentioned by someone else). Sorry I can't give more insight, but I can tell you this isn't Mercury crashing per se. Cheers! -- David --

HI, pls let me know how to send here a picture with a screenshot including the error message. on this picture is easy to see in the toolbar the mercury entry with the word 'error' inside.

when this message is accepted the mercury system restart. 

 

 

 

<P>HI, pls let me know how to send here a picture with a screenshot including the error message. on this picture is easy to see in the toolbar the mercury entry with the word 'error' inside.</P> <P>when this message is accepted the mercury system restart. </P> <P mce_keep="true"> </P> <P mce_keep="true"> </P> <P mce_keep="true"> </P>

If it's the same error message you quoted earlier I don't think we need to see a screenshot. However, you can insert an image here by using the "Insert/Edit image" button in the message toolbar.

Do you have any daemons running? Did you try deactivating them to see if that makes any difference?

/Rolf 

<p>If it's the same error message you quoted earlier I don't think we need to see a screenshot. However, you can insert an image here by using the "Insert/Edit image" button in the message toolbar.</p><p>Do you have any daemons running? Did you try deactivating them to see if that makes any difference?</p><p>/Rolf </p>

I think is important to show you the picture because it have evidences about mercury showing the word 'error' on the screen but i don't find the way to upload it because the image upload button on the tolbar need a url and the picture is located on my computer.

I think is important to show you the picture because it have evidences about mercury showing the word 'error' on the screen but i don't find the way to upload it because the image upload button on the tolbar need a url and the picture is located on my computer.

No one doubts that you get that error message, but as David explained it doesn't originate from Mercury itself. It could propagate from a daemon, though, so that's why I asked about that.

/Rolf 

<p>No one doubts that you get that error message, but as David explained it doesn't originate from Mercury itself. It could propagate from a daemon, though, so that's why I asked about that.</p><p>/Rolf </p>

[quote user="David Harris"]
It *might* be SpamHalter, but only Lukas would know that - I have no way of helping you with that. It might also be something else in the system, such as STunnel (as mentioned by someone else).
[/quote]

All my daemons have own exception handler what writing all exceptions to log file, if it is possible. So, it is not error in some my daemon, very probably. (Of course, maybe is something very broken...)

BTW: Runtime error 216 is "General protection fault"

So, check, if all your executables or DLL's are fine and is not broken.

<p>[quote user="David Harris"] It *might* be SpamHalter, but only Lukas would know that - I have no way of helping you with that. It might also be something else in the system, such as STunnel (as mentioned by someone else). [/quote]</p><p>All my daemons have own exception handler what writing all exceptions to log file, if it is possible. So, it is not error in some my daemon, very probably. (Of course, maybe is something very broken...)</p><p>BTW: Runtime error 216 is "General protection fault"</p><p>So, check, if all your executables or DLL's are fine and is not broken. </p>

I understand that it is a problem very difficult to debug, mainly when different software creators with different errors management concepts attend, also it is not too frequent and not easy to reproduce error.

the real problem when it happens is that the mercury loader doesn't continue with te mercury reload until someone presses on the 'accept' button on the error message screen. It is possible to have some way where the error message automatically be accepted and the loader continue with the mercury reload?

<P>I understand that it is a problem very difficult to debug, mainly when different software creators with different errors management concepts attend, also it is not too frequent and not easy to reproduce error.</P> <P>the real problem when it happens is that the mercury loader doesn't continue with te mercury reload until someone presses on the 'accept' button on the error message screen. It is possible to have some way where the error message automatically be accepted and the loader continue with the mercury reload?</P>

From the system's (and the loader's) point of view Mercury is still running until the error message button has been clicked, making the process terminate. Unless we can replicate the problem it would be difficult to catch this. Besides that it's of course better to take care of the real problem instead of doing a work-around, so it would be valuable if you could make some tests running Mercury with and without any daemons and external processes that interact closely with the program in your system.

/Rolf 

<p>From the system's (and the loader's) point of view Mercury is still running until the error message button has been clicked, making the process terminate. Unless we can replicate the problem it would be difficult to catch this. Besides that it's of course better to take care of the real problem instead of doing a work-around, so it would be valuable if you could make some tests running Mercury with and without any daemons and external processes that interact closely with the program in your system.</p><p>/Rolf </p>

the problem for me is this is a production system and not possible to made test on it, over the test machine (exactly the same hardware and a ghost copy of the whole production system)  i don't find a way to reproduce this error but be sure than when it occours the mercury system STOP works until someone click over the accept button and this is the real problem because the mercury does not work during this periode and all services are unavailables.

 

<P>the problem for me is this is a production system and not possible to made test on it, over the test machine (exactly the same hardware and a ghost copy of the whole production system)  i don't find a way to reproduce this error but be sure than when it occours the mercury system STOP works until someone click over the accept button and this is the real problem because the mercury does not work during this periode and all services are unavailables.</P> <P mce_keep="true"> </P>

If you pause all Mercury modules, you can scan the entire production system with f.ex. latest NOD32 and latest AdAware. I'f they do come up clean, you might have some form of hardware fault, either with your memory or disk drives. If you'd be having memory or nic failures, you'd normally get blue screens so, best guess is then the drives. We use SpinRite to deep dive into a drive, and CnW Recovery whenever something is broken with discs. A good look at the event logs also reveals anomalies with a system. But, always make backups before scanning the likes - since you never know when the s* hits the fan...

If you pause all Mercury modules, you can scan the entire production system with f.ex. latest NOD32 and latest AdAware. I'f they do come up clean, you might have some form of hardware fault, either with your memory or disk drives. If you'd be having memory or nic failures, you'd normally get blue screens so, best guess is then the drives. We use SpinRite to deep dive into a drive, and CnW Recovery whenever something is broken with discs. A good look at the event logs also reveals anomalies with a system. But, always make backups before scanning the likes - since you never know when the s* hits the fan...

i changed the way how the mercury start intalling a NTWrapper and running the mercury.exe as a service.

 

since this i receive every day/two days the message:

 

Informando de errores en cola: mercury.exe, versión: 4.6.2.0, módulo con error: mercuryi.dll, versión 4.6.2.0, dirección de error 0x0001008b.

translated: queue errors information: mercury.exe version 4.6.2.0 faulty module mercuryi.dll version 4.6.2.0, error address 0X0001008b.

 

 

<P>i changed the way how the mercury start intalling a NTWrapper and running the mercury.exe as a service. </P> <P mce_keep="true"> </P> <P>since this i receive every day/two days the message:</P> <P mce_keep="true"> </P> <P>Informando de errores en cola: mercury.exe, versión: 4.6.2.0, módulo con error: mercuryi.dll, versión 4.6.2.0, dirección de error 0x0001008b.</P> <P>translated: queue errors information: mercury.exe version 4.6.2.0 faulty module mercuryi.dll version 4.6.2.0, error address 0X0001008b.</P> <P mce_keep="true"> </P> <P mce_keep="true"> </P>

That would indicate that the problem is connected to the IMAP module. Is this where you use stunnel?

/Rolf 

<p>That would indicate that the problem is connected to the IMAP module. Is this where you use stunnel?</p><p>/Rolf </p>

yes, we use stunnel with IMAP4, POP3 and SMTP also with many others ports....

yes, we use stunnel with IMAP4, POP3 and SMTP also with many others ports....
live preview
enter atleast 10 characters
WARNING: You mentioned %MENTIONS%, but they cannot see this message and will not be notified
Saving...
Saved
With selected deselect posts show selected posts
All posts under this topic will be deleted ?
Pending draft ... Click to resume editing
Discard draft