D
david.karr
(I would post this on the Eclipse newsgroups, but some sort of bug in
the web interface is preventing me from logging in to post.)
Several weeks ago, I installed Eclipse 3.5 on a box for the first
time, and it was working perfectly fine. Last week, I installed it
again on a new box (different company), and it was also working fine,
for a while. At the end of last week, some event or threshold was
reached that I'm not aware of, such that it often fails with "out of
permgen space" errors, often when I'm viewing XML files more than a
few pages long. I've never had this kind of problem with Eclipse
before.
I'm currently sending "-vmargs -Xms64m -Xmx1152m" at the end of the
Eclipse command line, and I changed the two "--launcher.XXMaxPermSize"
options in eclipse.ini such that the passed value is "512m" (I imagine
setting both is redundant, but I wanted to make sure). Even with this
change, it still runs out of permgen space, along with numerous
"unhandled event loop exception" messages in the log.
I'm going to run VisualVM to watch the process for a while, to see
what it thinks is being used in permgen the next time it fails.
Any ideas about anything I can do to mitigate this or get more useful
information?
the web interface is preventing me from logging in to post.)
Several weeks ago, I installed Eclipse 3.5 on a box for the first
time, and it was working perfectly fine. Last week, I installed it
again on a new box (different company), and it was also working fine,
for a while. At the end of last week, some event or threshold was
reached that I'm not aware of, such that it often fails with "out of
permgen space" errors, often when I'm viewing XML files more than a
few pages long. I've never had this kind of problem with Eclipse
before.
I'm currently sending "-vmargs -Xms64m -Xmx1152m" at the end of the
Eclipse command line, and I changed the two "--launcher.XXMaxPermSize"
options in eclipse.ini such that the passed value is "512m" (I imagine
setting both is redundant, but I wanted to make sure). Even with this
change, it still runs out of permgen space, along with numerous
"unhandled event loop exception" messages in the log.
I'm going to run VisualVM to watch the process for a while, to see
what it thinks is being used in permgen the next time it fails.
Any ideas about anything I can do to mitigate this or get more useful
information?