Pete, thanks a lot. I installed from the master branch of github and the
problem was instantly resolved, so it was probably from that commit or
another that has since fixed the problem.
On Thu, Jan 2, 2014 at 8:07 PM, Pete Zaitcev wrote:
> On Mon, 23 Dec 2013 12:13:13 -0800
> Stephen Wood wrot
On Mon, 23 Dec 2013 12:13:13 -0800
Stephen Wood wrote:
> During any given 100 uploads, I'll usually see between 1 to 4% of the
> calls give a 404.
In my practice, all such cases were related to custer health. However,
since your 404 occurs when Object controler in proxy checks if container
exist
DB locked stack trace is a known issue.
How fast did you push 100 objects and what's your HW spec there ?
You can check for the existence of 404 container by swift-get-node.
In case of :
Dec 23 08:49:11 storage1 container-server 10.xx.xx.xxx - -
[23/Dec/2013:08:49:11 +] "HEAD /slot-8/11126
I'm also seeing this in the object-server logs:
12OperationalError: database is locked#012 (txn:
tx37d4e1807957447cac403-0052b89cbb)
Dec 23 20:27:39 store01 container-server ERROR __call__ error with PUT
/slot-3/11126/AUTH_swift/test1 : #012Traceback (most recent call last):#012
File "/usr/lib/py
