Beta after 747 have memory issues?

[browned]

Just a quick note to say that any beta I have tried after 747 has after a few hours caused the server to die a terrible and slow death of memory exhaustion.

 

I do not have logs and have not bothered to mention it before as I have been to busy. But on trying most builds after 747 to see if the performance and memory issues are fixed I thought I better post since I had some time.

 

I have VMWare based Windows 2016, 4 Cores, 8GB RAM (now 12GB), 2 x HPE P410 smart array cards with 7 x raid 0 arrays. The I have a pool within a pool, so Pool A consists of 7 x Raid 0 disks 20TB, Pool B consists of 7 x Raid 0 Disks 20TB, Pool C consists of Pool A and Pool B duplicated.

 

One thing I noticed is that stablebit was checking Pool C and it was at 3% for many hours, file details changing every 5 to 10 seconds.

 

Hopefully you have noticed something similar, or can replicate without logs as the impact on my server is too drastic for it to run more than a few hours.

[ZXTT95]

Just a comment about having no paging file.  This is equivalent to reducing your amount of RAM.  If you have plenty of RAM like you do, Windows won't be paging for the heck of it, but will use the paging file to store data that hasn't been accessed for a while and quite likely won't be accessed again as long as Windows is up and running.  The RAM that would be used by unused data can instead be used for things like caching.

 

At 64 GB, this isn't likely to make a difference, although as you go down to 8 GB and below, then disabling the paging file is likely to reduce performance.  The funny thing is if you go back to the days of home computers before virtual memory, people were constantly begging for it.

[Christopher (Drashna)]

Also, disabling the paging file will also prevent things like writing out crash dumps.   Without the paging file, it simply won't write them out.

[thepregnantgod]

Regarding the paging file, I suspected as much so I enabled a 64gb paging file on my Samsung 950Pro nVME.

 

The system RAM still crept up and up until lock.

[Christopher (Drashna)]

Regarding the paging file, I suspected as much so I enabled a 64gb paging file on my Samsung 950Pro nVME.

 

The system RAM still crept up and up until lock.

 

well, it wouldn't stop the memory leak, but it would allow it to be properly dumped, in this case.

 

But as gj80 mentioned, try switching away from ReFS and see if that helps.

 

Otherwise, enable pool tagging and use poolmon.

That may at least indicate what is eating up the RAM, kernel-wise.

https://support.microsoft.com/en-us/help/177415/how-to-use-memory-pool-monitor-poolmon.exe-to-troubleshoot-kernel-mode

 

 

However, poolmon is part of the Windows Driver Kit, and would require you to download the kit to gain access to the tools. 

https://developer.microsoft.com/en-us/windows/hardware/windows-driver-kit

[thepregnantgod]

Following your instructions...

 

I've added the bugbheck on registry.

I have a 64gb page file enabled (no ramdisk)

 

I hold the rightmost Ctrl key and hit scroll lock twice.  

 

Nothing.

[Christopher (Drashna)]

Following your instructions...

 

I've added the bugbheck on registry.

I have a 64gb page file enabled (no ramdisk)

 

I hold the rightmost Ctrl key and hit scroll lock twice.  

 

Nothing.

 

......

 

Yeah, I was just talking to alex about this, and ... I'm pretty sure that I've been experiencing the same thing. Especially as the hotkeys don't work for me when this happens. 

 

 

Well, if this is the same thing, I'm setup to get a dump "manually", so no need to do so.

But if you want.... download the Windows SDK, and install only the "Debugging tools" option.  then grab the "livekd" util for systernals.

 

From there, when you notice high usage, run "livekd -o memory.dmp -ml". 

Compress the dump file, and then upload it.

https://www.dropbox.com/request/gQCQNEeo7koiaxLpBhgq

 

Either way, this sounds more and more like ReFS isn't ready for prime time, as it does have issues.

[thepregnantgod]

So, am I reading, Chris, that you're experiencing what I'm experiencing?  As I remove my REFS drives from my pool and I'm noticing the memleak isn't hitting as high.  Right now it's stopping around 30gb RAM instead of maxing at 64gb and softlocking my system (I can move my mouse but not select anything).

[Christopher (Drashna)]

Yes, but nowhere near as badly as you.  High memory usage from my system, but not exhausting the memory.  I have 16GB in the system, and I see 8-10GBs used. However, rarely, I do see it on the high side (14+GBs), and that's when "weird things happen".  And occasionally, that's a black screen... where the keyboard command doesn't work. 

 

That said, I recently reimaged the system to an SSD (long story) and that has helped, but I'm sure that's in part because I do have a large pagefile.

 

All in all, not enough to point to a memory leak, until you started talking about your system.   So... it definitely sounds like ReFS has some issues. Unfortunately. 

[thepregnantgod]

I'm hoping to have the last of my REFS drives removed from the pool this weekend.

 

I plan on reinstalling Windows, only my essential programs (chrome, norton, etc.) then drivepool.

 

I'll build a new pool with the evacuated drives and start copying my files over there and track memusage.

 

Really don't want to go to Storage Spaces because I've been burned by RAID before but I can't have my system soft-lock every 6hrs either.  (And upgrading to 128gb RAM is too expensive)

[Christopher (Drashna)]

Well, let me/us know how it goes.  

 

Hopefully, this does fix the issue, though .... that means no ReFS, which is just as disheartening. 

 

And I totally understand not wanting to use Storage Spaces.  And yeah, that much RAM is expensive.