On Jan 11 2017, Chris Davies <[email protected]> wrote:
> I have two large (multi-TB) filesystems running on S3QL. One of them 
> contains a number of large files (several hundred GB). Unfortunately the 
> filesystem has crashed a number of times over the last few months. Recently 
> these crashes are becoming more and more frequent - several times a day. I 
> think I've finally identified the cause, although I don't have a good 
> resolution.
>
> It seems that if the filesystem discovers that it is deleting a file, but a 
> block supposedly belonging to that file does not exist, the filesystem 
> crashes in a heap. Normally I'd say that the correct resolution would be to 
> run the filesystem checker in full verification mode, and maybe that's 
> where I'll end up heading if I can't get a better fix.

What backend are you using? For most sane backends, a regular fsck.s3ql
should fix these issues.

> However, in this 
> particular use case I don't think the filesystem should crash.

You are right, and it actually should not do that but enter "failsafe"
mode instead (file system becomes read-only). Could you please file an
issue at https://bitbucket.org/nikratio/s3ql/issues? I'll try to get
this fixed in the next release.


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.«

-- 
You received this message because you are subscribed to the Google Groups 
"s3ql" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to [email protected].
For more options, visit https://groups.google.com/d/optout.

Reply via email to