[zfs-discuss] Issue #7204: Question about "refcount mismatch"

Matthias Bodenbinder matthias at bodenbinder.de
Sun Feb 25 03:03:13 EST 2018


the issue #7204 I reported on github (https://github.com/zfsonlinux/zfs/issues/7204) was "locked as off-topic and limited conversation to collaborators" by user @kpande. He suggested that I seek happiness in this user group. 

My question is: What is the risk associated with the zdb warning (error?) message: "space map refcount mismatch: expected 240 != actual 232" ? 

This issue was first reported more than 3 years ago and I can not find a commonly accepted risk assessment of this issue. On github I found several related issues with comments from developers which are partially contradicting each other. On zfs-discuss I did not find any posting related to this topic.

User @kpande on github is saying: "this topic has been beaten to death ... for further reading see #3094". I could not agree less. When searching the web I can find user reports and users opinions, but nothing really solid. And Issue #3094 "Spacemap Problem with Data Error" contains no risk assessment about this issue. I have known #3094 before I created #7204. But #3094 is too complex and incomprehensible for simple users like me. Thats why I created #7204 in the first place - as a "Question". But somehow this did not work. The issue was closed and now I am hear. 

What is the risk for my data, my pool if I see a message like "space map refcount mismatch: expected 240 != actual 232" ? 


More information about the zfs-discuss mailing list