Quantcast

Severe performance hit over time with mmap usage

classic Classic list List threaded Threaded
1 message Options
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Severe performance hit over time with mmap usage

Van Breuk
Dear devs,

I am testing Mergerfs to bundle two SSD drives, in order to handle a large database that was surpassing the single drive capacity. This database is run by a process that uses extensively large memory mapped files, on a high RAM (64 GB) Ubuntu server.

After reading in the Mergerfs documentation https://github.com/trapexit/mergerfs/blob/master/README.md about FUSE performance issues when running mmap with older kernels, I upgraded my server to 4.5.4 kernel.

Still, the process starts fine from the Mergerfs mount point, with very reasonable resource usage, but after a few hours Mergerfs is staying over ~100% CPU usage, and the whole system lags noticeably.

Trapexit, developer for Mergerfs, suggested to bring it here since it could be essentially a FUSE issue, perhaps to do with caching. I will probably switch to another device management scheme like LVM, but if there's any details I can provide about my issue, please let me know.

Regards,

Van


------------------------------------------------------------------------------
Check out the vibrant tech community on one of the world's most
engaging tech sites, SlashDot.org! http://sdm.link/slashdot
--
fuse-devel mailing list
To unsubscribe or subscribe, visit https://lists.sourceforge.net/lists/listinfo/fuse-devel
Loading...