-
1. Re: Clusted Cache with Berkeley CacheLoader with a singleton
manik Nov 20, 2007 4:57 AM (in response to jivesociety)Is C:\cache\stats on a shared drive? Or perhaps a previous process did not end correctly and left a stale je.lck lock file?
-
2. Re: Clusted Cache with Berkeley CacheLoader with a singleton
jivesociety Nov 20, 2007 11:52 AM (in response to jivesociety)C:\cache\stats is not a shared drive. But, I was contemplating going that route as we are planning on deploying this to unix systems with a shared NFS filesystem.
It seems that the first process in the cluster creates the je.lck file to signify it is the writer process. And every other node that starts up afterwards throws this error due to that lock file being present.
I might have misconfigured the CacheLoaderConfig, is this configuration allowable though?
thanks in advance,
ray -
3. Re: Clusted Cache with Berkeley CacheLoader with a singleton
manik Nov 27, 2007 7:22 AM (in response to jivesociety)Read http://www.oracle.com/technology/products/berkeley-db/faq/je_faq.html#1 before you think about BDBJE on NFS! :-)
It may just be that BDBJE in a shared environment may not work due to the way the BDBJE engine implements locks.
Have you tried JDBM instead of BDBJE?