-
1. Re: Frequent
rob.stryker Jul 5, 2006 6:09 PM (in response to znbailey)The only thing I can think of for that error is if the server is taking longer than my defined timeout.
How long is it taking your server to start? If it is actually close to where my timeout is, I can either try to increase the timeout or allow the user to do it. -
2. Re: Frequent
znbailey Jul 5, 2006 6:21 PM (in response to znbailey)Rob,
It would be great if that were a user-configurable option. We have a considerable deployment, such that when debugging it is not uncommon for it to take anywhere from 55s to 1:35s to fully start up. (this is according to the line that says "17:57:10,562 INFO [Server] JBoss (MX MicroKernel) [3.2.7 (build: CVSTag=JBoss_3_2_7 date=200501280217)] Started in 1m:10s:328ms", for example). -
3. Re: Frequent
jiyanghu Jul 6, 2006 2:02 PM (in response to znbailey)I got the same issue - failed to start without any info(I am using 3.2.8sp1). It stopped starting after 2 minutes based on console log info. No reason for such a failure. I used the same ear file with IDE1.6 and it starts fine.
-
4. Re: Frequent
znbailey Jul 6, 2006 2:13 PM (in response to znbailey)Rob,
Any word on this? What is the timeout you have defined? What would be some other reasons for getting this "failed to start" message? It seems that maybe the 2.0 IDE package is less forgiving when certain non-critical startup items fail (i.e. failures not reported to the console, but may be happening internally during the JBoss AS startup process, which do not harm the startup process). -
5. Re: Frequent
znbailey Jul 6, 2006 3:13 PM (in response to znbailey)Here is the result of copying the event details to the clipboard from the error log:
eclipse.buildId=M20060629-1905 java.version=1.5.0_05 java.vendor=Sun Microsystems Inc. BootLoader constants: OS=win32, ARCH=x86, WS=win32, NL=en_US Command-line arguments: -os win32 -ws win32 -arch x86 Error Thu Jul 06 12:15:58 EDT 2006 Server JbossServer Cascade Server 4.x (TRUNK) failed to start.
In the error log it says this happened in plug-in "org.eclipse.wst.server.core". -
6. Re: Frequent
znbailey Jul 11, 2006 12:26 PM (in response to znbailey)A little more information:
This problem appears to crop up a lot more frequently (pretty much 100% of the time I get the "failed to start message") when I am doing other things while I am waiting for the server to start (both things inside of and outside of eclipse). I am on a dual-processor machine. -
7. Re: Frequent
rob.stryker Jul 11, 2006 12:31 PM (in response to znbailey)For this week's release I've up'd the default timeout to 180000 ms, and I've added what I feel is better information to the event log regarding what's going on. You'll also be able to select an item in the 'event log' section of the view and, if the properties view is visible, get some additional information.
Hopefully this will help. At least I'm certain if it happens again I'll get more useful information back.
- Stryker -
8. Re: Frequent
znbailey Jul 12, 2006 12:01 PM (in response to znbailey)"rob.stryker@jboss.com" wrote:
For this week's release I've up'd the default timeout to 180000 ms, and I've added what I feel is better information to the event log regarding what's going on. You'll also be able to select an item in the 'event log' section of the view and, if the properties view is visible, get some additional information.
Hopefully this will help. At least I'm certain if it happens again I'll get more useful information back.
- Stryker
Rob,
Sounds great. Thanks for all the hard work. -
9. Re: Frequent
mjdinsmore Jul 12, 2006 11:11 PM (in response to znbailey)Rob, You mentioned:
"rob.stryker@jboss.com" wrote:
How long is it taking your server to start? If it is actually close to where my timeout is, I can either try to increase the timeout or allow the user to do it.
Does your later comment:"rob.stryker@jboss.com" wrote:
For this week's release I've up'd the default timeout to 180000 ms, and I've added what I feel is better information to the event log regarding what's going on. You'll also be able to select an item in the 'event log' section of the view and, if the properties view is visible, get some additional information.
Hopefully this will help. At least I'm certain if it happens again I'll get more useful information back.
- Stryker
mean that you did not add the user configurable timeout? I tried changing...\eclipse\plugins\org.jboss.ide.eclipse.as.core_1.0.0.Alpha\plugin.xml
with the startTimeout and stopTimeout values to something else but it didn't seem to have any effect. I'd be happy if the timeout was set to 300000 ms as even 180000 ms isn't going to be anywhere close to long enough for things to start up... (the larger value would at least give me breathing room)
Am I missing something? Should that manual change in the plugin.xml file actually be used?
thanks,
Mike -
10. Re: Frequent
rob.stryker Jul 12, 2006 11:34 PM (in response to znbailey)Here's the issue: the plugin.xml requires a global max. For right now I set the global max at 7.5 minutes (450,000 ms). Any user-defined max will not be able to go beyond this global max. If you think this global max is too low, please suggest a higher one.
For the release we are making this week, there is no preference dialog yet, and you are correct to notice that changing the xml does not change the results for you.
My code checks the plugin preferences (which you are not yet able to set) and if they are not set, it defaults to 180000 ms (instead of the 120000 ms from before, but as you said still not high enough for you.)
The build this week will go out with this 3 minute effective max unless you all make a big stink about it and demand a higher one, in which case I can probably convince Marshall to initiate a rebuild and set the effective max to the global max for this release.
I suppose I should have done that anyway. -
11. Re: Frequent
znbailey Jul 13, 2006 10:28 AM (in response to znbailey)Rob,
That sounds good. I'm sure there may be people here whose servers take that long to start, but I think we can imagine they would be in the minority.
I will see if the new build addresses this issue. However I am not sure it was the timeout issue I was ever suffering from - I have seen this error screen pop up once before when my server took more than 2 minutes to start - and it was different than the error dialog I receive most of the time which is the generic "the server failed to start" rather than the "timeout exceeded" error message.
I will report back in this thread when the new release hits. Thanks! -
12. Re: Frequent
znbailey Jul 18, 2006 11:59 AM (in response to znbailey)I just upgraded to Alpha1a. Now, when starting the server the status box ALWAYS says "starting...", even after I have received the message "11:47:57,734 INFO [Server] JBoss (MX MicroKernel) [3.2.7 (build: CVSTag=JBoss_3_2_7 date=200501280217)] Started in 56s:922ms". Then, after a while the error dialog pops up : "The server failed to start in 450 seconds", but that error dialog does not make the server stop running. It seems there is some disconnect between the actual state of the server and the state returned by the method the jboss adapter uses to poll the server.
-
13. Re: Frequent
rob.stryker Jul 18, 2006 2:55 PM (in response to znbailey)znbailey:
could you please provide me with more information from the jboss server view's event log? Also, you can get some extra information by opening the properties view and clicking on various event log items to see some properties of them.
Please provide as much data from there as you can. You are correct. There is a disconnect of some sort. What type of configuration are you using? Is it a modified minimal configuration or a generic standard configuration?
What JBoss version are you using?
etc etc. -
14. Re: Frequent
znbailey Jul 18, 2006 4:12 PM (in response to znbailey)Rob,
Absolutely. My configuration information is as follows:
JBoss 3.2.7, using the "default" configuration. Deploying in a flat file EAR.
The error message in the error console (and in the dialog that appears) is as follows:Timeout waiting for JbossServer Cascade Server 4.x (TRUNK) to start. Server did not start after 450s.
The Exception Stack Trace is not available.
Here is the session data, as available from the "event details":eclipse.buildId=M20060629-1905 java.version=1.5.0_05 java.vendor=Sun Microsystems Inc. BootLoader constants: OS=win32, ARCH=x86, WS=win32, NL=en_US Command-line arguments: -os win32 -ws win32 -arch x86
My program arguments are : "--configuration=default --host=localhost"
VM arguments: "-Xms256m -Xmx512m"
Let me know if I can get you any other information.
Thanks.