Hi Phil, On Thu, Oct 13, 2011 at 9:50 AM, Phil Rice <[email protected]> wrote: > ...When I restart the server, instead of the usual "welcome to apache > sling" page, with the resource explorer etc, I had a directory listing > of the root of the server, with the text "powered by jackrabbit" at > the bottom. I was unable to get the usual sling page to appear...
Did you look at /system/console to see which bundles/components might have been down? > > Several server reboots later, with no configuration parameters > changing, the sling server decided to behave and I have the usual > welcome index page.. I see two possible reasons, either the JCR repository needed time to recover from the unexpected reboot (but I don't know exactly how Jackrabbit handles this, so just a guess) or the OSGi framework state under sling/launchpad was corrupted and somehow got back to life in one of the restarts (but I don't know if/how that could happen either). > > The main reason I am concerned is that I don't know what happened, and > should it happen again I have no idea where to begin troubleshooting > such a problem. This server is going live very soon, and I hope to > have hundreds of users accessing it. > > There was no smoking gun in the error logs, although there were a lot > of "dependency not found errors".... Which might indicate that the JCR repository was not running, and came back later. Might just be caused by a .lock file [1] left in the repository, but that shouldn't go away without your intervention AFAIK. [1] http://wiki.apache.org/jackrabbit/RepositoryLock > > So I guess the real questions I am asking are: > * Why did it show the "powered by jackrabbit" front page? I don't know, and I don't even know exactly which page that is, do you have a screenshot? > * How do I fix sling should this happen again? Did you save the state when your system was broken? If yes you can go back to it and investigate. If not, I'd try breaking it again by killing it repeatedly from a script until you can reproduce the problem. > * Are there any information pages about managing sling in a production > server that I should read? Not that I know of, but my guess is that what you saw is more related to the Jackrabbit or to the Felix OSGi framework, so not directly to Sling. The only way to find out is to reproduce the problem and dig deeper to find out exactly what was missing. -Bertrand
