Benjamin Arntzen

off-topic, the HC2 is an awesome unit for MooseFS/LizardFS. I've got a cluster of 4. Love them.

RC1 is buggy as hell. Sticking with it seems foolish. It has multiple EAT YOUR DATA BAD bugs.

Alternately, you could also consider RAM requirements as being twice the size of the metadata set. At my previous employer that was the answer - gobs of RAM with plenty...

+1 - my approach would be to rebrand every single user facing thing in one release (MFS -> LFS) with a notice that we will be breaking backwards compatibility for...

@blink69 What is the timeline for 5.x?

@guestisp you're missing chunks of the conversation because you blocked 4dolio. Maybe time to end that feud?

IIRC this is already on the roadmap. - b On 11/05/2015 1:01 pm, "Dmitry Smirnov" [email protected] wrote: > One of the LizardFS features we are badly missing is global cache...

It's got a single disk. I'll try that later tonight, thanks :) Might also try a bunch of other distributed FS just for comparison, despite my strong preference for Lizard...

This is absolutely a bottleneck somewhere in LizardFS. I downloaded and built from the latest source MooseFS (3.0.101), then created a chunkserver using the exact same XFS brick (down to...

Running a dd test on MooseFS: root@blinky:/mnt/mfs# dd if=/dev/zero of=./largefile bs=1M count=1024 1024+0 records in 1024+0 records out 1073741824 bytes (1.1 GB, 1.0 GiB) copied, 8.08992 s, 133 MB/s