Memory leak

[barbergeek]

Have a new problem that showed up with b12. There seems to be a continuous memory leak that eventually eats the entire available memory pool. I thought it was one of the Docker containers, but Docker is stopped and memory usage, according to the system status, continues to go down. Eventually it gets to 98% and then I lose access to the UI. After rebooting and watching it go up for several hours, I stopped and restarted the array and it went back down to about 10%. It's been sitting (without Docker running) for about 3 hours and it's up to 19%.

 

I have a pretty stock b12 build, with only dlandon's powerdown v2.13  plugin installed.

 

I have been watching htop too and it's gone in that same time from 252 to 454, which, given the 16GB I have installed isn't much, but eventually it gets yellow bars all the way across and a number up close to the 16 gig.

 

Anyone else seeing this or suggestions to help debug?

 

Thanks!

[BRiT]

This is not a memory leak. Read the following site linuxatemyram.com

 

[barbergeek]

Well, I am definitely losing access to the UI after about 24 hours or so. I will let everything run for a while and see what free -m reports. Definitely different behavior in b12 than what I had in b10, as it was running pretty great.

[barbergeek]

I guess I would add that if this really is the case, the system stats on the dashboard are pretty misleading (and I don't consider myself at least a total noob  ) and it would be helpful to break those stats down a bit into maybe used, cached, and available, or something such.

 

If the GUI continues to die on me, I'll try to debug further.

 

Thanks for the clarification. We'll see what happens.

[PeterB]

I guess I would add that if this really is the case, the system stats on the dashboard are pretty misleading (and I don't consider myself at least a total noob  ) and it would be helpful to break those stats down a bit into maybe used, cached, and available, or something such.

 

If the GUI continues to die on me, I'll try to debug further.

 

Thanks for the clarification. We'll see what happens.

 

If the gui dies and you still have console/teminal access, see whether you have two instances of /usr/local/sbin/emhttp active:

ps -eaf | grep emhttp

 

I'm trying to track this issue down, too.

 

Thread here.

 

 

[barbergeek]

I installed the Dynamix v6 system stats plugin and it was quite informative as to what's really going on with the memory. Might want to consider that as part of the standard install. No memory leak evident, definitely using lots of cache though, as it should, my apologies for the confusion.

 

I was occasionally still having difficulty getting to the GUI, but then I noticed it was only on my iPad. I cleared cache and have been unable to reproduce. Odd. I will monitor and opena new thread if I see it pop up again.

 

Thanks everyone for helping clarify.

[PeterB]

I installed the Dynamix v6 system stats plugin and it was quite informative as to what's really going on with the memory. Might want to consider that as part of the standard install. No memory leak evident, definitely using lots of cache though, as it should, my apologies for the confusion.

 

The problem, of course, is that what is displayed as 'allocated' memory on the dashboard should not include that memory allocated to the cache.  If the dashboard displayed the 'used' memory, as shown in the system stats plugin, then there would be little confusion and no cause for alarm.