severity 771452 important
thanks

On 12/02/2014 01:50 PM, Shannon Dealy wrote:
> On Mon, 1 Dec 2014, Nikolaus Rath wrote:
> 
>> severity -1 normal
>> retitle -1 fsck.s3ql sporadically crashes when checking objects
>> thanks
>>
>> Retitling this bug and lowering severity for now. I don't think that the
>> mount.s3ql crash is related to the fsck.s3ql crashes, or that there is
>> any data corruption.
> 
> While I understand your assessment, I don't think I would characterize
> this bug as "normal".  Based on my understanding, the Debian guidelines
> would rate this bug at the very least as "important", and frankly I
> would still argue for "critical" given that it effectively "causes
> serious data loss" (one of the possible criteria for "critical").
> 
> My reasoning is that even assuming the file system data has not actually
> been lost or corrupted (and I believe you are likely correct in this), I
> ran fsck.s3ql roughly 30 times before getting a successful run.  There
> is nothing to indicate that someone else with similar circumstances
> would not have to run it 1000 or 100,000 more times in order to recover
> the file system.  For such a case, their data is unavailable to them and
> effectively 100% lost until they have a fixed version of fsck.s3ql

I think the crucial difference is that the data is inaccessible, not
lost. It can easily be retrieved by upgrading S3QL, and chances are that
enabling/disabling SSL or using a different network connection may also
help. A data *loss* means that data is destroyed, not that it's
unavailable. Putting it differently, anything that is "lost until x" is
not lost.

But there's little point in arguing about severity. I'll set it to
important, that's still enough to get an update into jessie.


Best,
Nikolaus

-- 
GPG encrypted emails preferred. Key id: 0xD113FCAC3C4E599F
Fingerprint: ED31 791B 2C5C 1613 AF38 8B8A D113 FCAC 3C4E 599F

             »Time flies like an arrow, fruit flies like a Banana.«


-- 
To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org

Reply via email to