Re: Addressing OOM while recovering overflow regions.

2017-01-18 Thread Anilkumar Gingade
Thanks for all the comments... As suggested we will go with option#1 and additionally add a system property to support reading values for lru regions. -Anil. On Tue, Jan 17, 2017 at 9:07 PM, Nilkanth Patel wrote: > Anil, option 1 looks good. Just have a small

Re: Addressing OOM while recovering overflow regions.

2017-01-17 Thread Nilkanth Patel
Anil, option 1 looks good. Just have a small question. Will the option 1 implementation going to replace the existing implementation where we recover everything? Or will this be a default and still user will be able to use old/existing way based on some configuration where recovering everything.

Re: Addressing OOM while recovering overflow regions.

2017-01-17 Thread Michael Stolz
Can I assume all options load keys? If so, yes skip loading values. -- Mike Stolz Principal Engineer - Gemfire Product Manager Mobile: 631-835-4771 On Jan 17, 2017 5:48 PM, "Anilkumar Gingade" wrote: > Hi Geode Devs, > > We are working on ticket GEODE-1672, related to out

Re: Addressing OOM while recovering overflow regions.

2017-01-17 Thread Udo Kohlmeyer
I think that #1 is a good start... I think we should have all regions created before we recover from the files, or at least for the regions that use the diskstore being recovered. When we recover, do we start recovering in reverse? I mean, the latest data *should* be in the last data files,

Re: Addressing OOM while recovering overflow regions.

2017-01-17 Thread Jason Huynh
I like number 1 On Tue, Jan 17, 2017 at 3:56 PM Dan Smith wrote: > Option 1 is not a bad idea. Another thought might be to not start > asynchronous value recovery until all of the regions are created. I think > right now we launch a task to read all of the oplogs and recover

Re: Addressing OOM while recovering overflow regions.

2017-01-17 Thread Dan Smith
Option 1 is not a bad idea. Another thought might be to not start asynchronous value recovery until all of the regions are created. I think right now we launch a task to read all of the oplogs and recover values as soon as the disk store is created. Maybe we could defer that until after the last

Re: Addressing OOM while recovering overflow regions.

2017-01-17 Thread Hitesh Khamesra
option 1 looks good to me.. From: Anilkumar Gingade <aging...@pivotal.io> To: dev@geode.apache.org Sent: Tuesday, January 17, 2017 2:41 PM Subject: Addressing OOM while recovering overflow regions. Hi Geode Devs, We are working on ticket GEODE-1672, related to out of memory