[sheepdog-users] What can I do with my .stale objects ?

[Hitoshi Mitake]

At Sun, 11 Jan 2015 01:33:36 +0100 (CET),
Walid Moghrabi wrote:
> 
> [1  <multipart/alternative (7bit)>]
> [1.1  <text/plain; utf-8 (7bit)>]
> Hi, 
> 
> 
> I had some troubles recently with my 2 node cluster. 
> I'm running 0.8.2 and so far so good except that, due to my network configuration, I must run corosync using unicast udp. 
> While doing a massive rsync between one of my nodes local filesystem and one of my VDI stored on Sheepdog, I encountered i/o problems due to network overload which made Corosync go into complete mess. 
> Once I understood the problem, I took measures to limit bandwidth pressure and checked/repaired every VDIs (or delete them when needed), I ended up with a fully functionnal Sheepdog cluster with VDIs in good shape except that, on one of my 2 nodes, I have the /var/lib/sheepdog/obj/.stale which is full of data (about 1.7TB) which I don't have on the other node. 
> I tried dog vdi cache purge but that didn't clean anything. 
> What Can I do with this ? 

What is redundancy of your VDIs? If it is 2, and your VDIs seem to be healthy, you can remove them safely.

> Can I simply discard these files manually and is it a safe procedure or is there any better way to do this ? 

Remaining stale objects would be a bug of sheepdog, could you provide log of the node?

Thanks,
Hitoshi

> 
> 
> Best regards, 
> 
> 
> Walid Moghrabi 
> [1.2  <text/html; utf-8 (quoted-printable)>]
> 
> [2  <text/plain; us-ascii (7bit)>]
> -- 
> sheepdog-users mailing lists
> sheepdog-users at lists.wpkg.org
> https://lists.wpkg.org/mailman/listinfo/sheepdog-users