drbd_unsynced multiple is wrong

While setting up some new drbd pairs I’ve noticed that the ‘drbd_unsynced’ amount is out by a factor of *0.01 - for example it’s showing me 300M of unsync’d data when it is actually 30,000M / 30G

To be certain, I created a 40G test volume and stopped it from synching for a bit so I could get some data.

Grabbed the output of /proc/drbd just after the graph rendered:

1001: cs:WFConnection ro:Primary/Unknown ds:UpToDate/Inconsistent C r-----
ns:276640 nr:0 dw:42078576 dr:54061469 al:200 bm:0 lo:0 pe:0 ua:0 ap:0 ep:1 wo:f oos:29825160

The graph from the same time is attached.

The volume is new, 40G in size and has just had 40G of random data written to it on the primary. The secondary was deliberately taken offline after about 4G was sync’d (10%).

Just before taking it offline, I grabbed the resync data which interestingly uses Mbytes (not bytes) for the percentage remaining data – consistent of them, not.

1001: cs:SyncSource ro:Primary/Secondary ds:UpToDate/Inconsistent C r-----
ns:270496 nr:0 dw:42078576 dr:46059913 al:200 bm:0 lo:0 pe:0 ua:0 ap:0 ep:1 wo:f oos:37826716
[>...................] sync'ed: 9.9% (36940/40960)M
finish: 0:15:17 speed: 41,224 (23,256) K/sec

So, I suspect this has contributed to the wrong multiple factor being applied to the ‘oos’ element in the import into Observium.

If any additional info is required please do let me know!