I'm using Mach-II on several projects, some of them running under
railo (the latest version) without any problems.
On one setup (a quite small project) I receive the following error
message very often, however:
there is a timeout occurred on a exclusive lock with name
[_MachIILogFactory.channel_9E03698DF4712F7D0EBBB1F8BF991D6C] after 10
seconds
/var/www/MachII/logging/LogFactory.cfc (68)
/var/www/MachII/framework/ViewContext.cfc (238)
/var/www/MachII/framework/ViewContext.cfc (46)
/var/www/MachII/framework/EventContext.cfc (78)
/var/www/MachII/framework/RequestHandler.cfc (195)
/var/www/MachII/framework/RequestHandler.cfc (130)
/var/www/MachII/mach-ii.cfc (153)
/var/www/processing/Application.cfc (90)
The error occurs with both 1.6.1 and the RC of 1.8
Running on Railo 3.1.2.008 final and 1.6.0_17 (Sun Microsystems Inc.)
Best regards,
Hansjoerg
I believe this is an issue with Railo directly. The reason why I'm
going in this direction is this first time I've heard of this issue and
I know for sure I've never seen this exception on OpenBD or Adobe CF.
We don't officially support Railo yet for Mach-II because there has been
some issue with the engine in the past. However that doesn't mean that
people aren't successfully using Railo in production. I wonder if there
is an issue with Railo.
* Are you seeing long on application start times for this application?
Maybe change the timeout from 10 to 30 and try again. Maybe for some
reason in your situation Railo is taking a long time creating that
Log.cfc. The Log.cfc is very lightweight so I would be surprised it
would take 10 seconds to do. There could be an issue with Railo and
releasing locks as well.
Let's figure out what's going on.
Best,
.Peter
Hansjoerg said the following on 12/02/10 02:59:
> I believe this is an issue with Railo directly. The reason why I'm
> going in this direction is this first time I've heard of this issue and
> I know for sure I've never seen this exception on OpenBD or Adobe CF.
The interesting thing is that this happens on a low-traffic engine;
there is no such issue with a high-traffic engine.
> * Are you seeing long on application start times for this application?
Not at all ...
>
> Maybe change the timeout from 10 to 30 and try again. Maybe for some
> reason in your situation Railo is taking a long time creating that
> Log.cfc. The Log.cfc is very lightweight so I would be surprised it
> would take 10 seconds to do. There could be an issue with Railo and
> releasing locks as well.
>
I've now increased the timeout to 30 seconds and will monitor the
engines.
Best regards,
Hansjoerg
FYI, I have filed a ticket for the issue you mentioned after I
reviewed the code more directly.
http://greatbiztoolsllc.trac.cvsdude.com/mach-ii/ticket/556
I believe the problem is that exclusive named locks in all CFML
engines are global across all applications on the same CFML server.
Although the lock name seems unique (e.g.
MachIILogFactory.channel_9E03698DF4712F7D0EBBB1F8BF991D6C), the hashed
part is the "channel" name of the logger. There are a lot of common
channel names in Mach-II applications like MachII.framework.*. I
believe the issue is a concurrency issue with another application on
the server starting up.
This will be fixed in Mach-II Integrity (1.9) and possibly backported
to the 1.8.1 maintenance release.
Best,
Peter
Team Mach-II
On Feb 14, 4:33 pm, Hansjoerg <off...@tunesbag.com> wrote:
> Hi Peter -
>
> > I believe this is an issue withRailodirectly. The reason why I'm
> > going in this direction is this first time I've heard of this issue and
> > I know for sure I've never seen this exception on OpenBD or Adobe CF.
>
> The interesting thing is that this happens on a low-traffic engine;
> there is no such issue with a high-traffic engine.
>
> > * Are you seeing long on application start times for this application?
>
> Not at all ...
>
>
>
> > Maybe change the timeout from 10 to 30 and try again. Maybe for some
> > reason in your situationRailois taking a long time creating that