Sambamba depth base does not accurately list depth per base in output

[Sven Winter]

Hi, I have tried several older and newer versions of Sambamba but for whatever reason my output list the header and then starts listing each position of the genome up to a random point and then prints the header again and starts again at the first base position. The last base before it goes back to the start seems to be random and and changes every time. 

Any Idea why this is? I am running a conda installation on a large HPC cluster with a job array to run each chromosome of my assembly in its own job.  The bam files I use are 46 cleaned bams with unmapped reads, repetitive regions, and smaller scaffolds removed. 

Any idea why it is doing that? 

I did use Sambamba on a different server before and it worked just fine and gave one line per base position.

Any help is appreciated!

Sven