We are currently preparing to go live with an app based on the 0.9.5 glassfish gem. We use 3 windows server hosts with on each host a non-dynamic pool of 4 glassfish instances. Each host uses the same shared jruby and app directories.
Today we experienced a logfile explosion. Almost every second a logfile was being created of about 2 MB. The growth rate was near 10GB/hour The content was almost in all files identical to the original log file.
It looked like the files were typically being created by a rotation task, they had a timestamp in their name and no .log suffix.
The weird thing is I reported a not functioning logfile rotation task a while ago (http://rubyforge.org/tracker/index.php?func=detail&aid=26393&group_id=5450&atid=21080)
I restarted the servers and deleted the tmp/.glassfish directory and so far this problem hasn't occurred again, although the log file hasn't reached the 2MB limit yet.
Perhaps the two problems are related?
Alex
|
Start New Thread
logfile explosion [ reply ]