I thought it was odd that our server always stayed shy of 8GB, despite -Xmx=16g ...but that was with version 0.197.51.
Now, the server just stuffs its face, with version 0.198.131
Originally wrote this off as an odd quirk the first few times...but there's a pattern here: ballooning has been occurring since the upgrade to 0.198.131. What additional information can we provide?
<replace this line with the file content>
<replace this line with the file content>
-Validated-
It usually doesn't go over the initial memory unless it's needed, so that's why it stayed under 8GB since the GC cleaned up everything nicely.
Doesn't seem to be the case anymore.
I assume that this was in a timespan of a few hours? Or is there a day or more between both screenshots?
I agree, something became really greedy between the transition of the 2 versions. On my side my connection has improved (download and upload speed improved) and nevertheless my localhosted have really hard time to be fluid (with the same settings ping pikes occures really often compare to the 197 version).
So even if it's a rejected one I put the link for archive here, it might help: T1461 - 0.197.52 to 0.198.131 same settings but not same lags.
Ahh, I see what you were asking. We normally auto-restart every 24hrs, which hasn’t been a problem until now. We might shift to 12hr restarts until this issue is addressed.
To clarify my original posting though: the first screenshot was from before the updates a few days ago. The 2nd screenshot was an hour or two prior to opening this ticket. Normally, for this "time of day" memory would only read at ~8GB consumed.
Putting in queue, did a heapdump on test server.
It's name: heapdump-24052016.tar in tmp folder
This task is too general since it's about no specific leak anymore. Many leaks have been fixed.