[tahoe-dev] 1.6.0 When storing shares (from another machine), Python consumes --> 100% of local (storage node) CPU

[Jody Harris]

On Sat, Mar 6, 2010 at 2:20 PM, Zooko Wilcox-O'Hearn <zooko at zooko.com>wrote:

> Thank you for the problem reports.
>
> On Saturday, 2010-03-06, at 12:55 , Jody Harris wrote:
>
> > My remote server is storing a "largish" file. On my notebook
> > computer (running 1.6.0) Python is consuming --> 100% CPU (of one
> > processor). My desktop computer (running 1.5.0) is not showing the
> > same symptoms at this time.
>
> I don't fully understand what's happening. Are there three machines
> involved -- your  server, your notebook, and your desktop? What
> operation are you doing -- an upload? Which machine is the web
> gateway (== storage client) running on when it does this upload?
>

"Server":
  - Web server hosted at Rackspace.com
  - storage operation running to grid on this machine
  - no storage nodes running on this machine
  - running 1.5.0

"Cat":
  - Desktop machine
  - running 1.5.0
  - storage node only
  - no symptoms indicated during the share store when "Ricki" was running
"hot"

"Ricki":
  - notebook computer
  - running 1.6.0 (modified during D-S's sftp hacks)
  - storage node
  - Machine that is running "hot" during share stores.


> > I attempted to run backups against the Volunteer grid for a week,
> > and experienced failures every night. One file or another (often
> > several files) would just stall completely. I switched back to the
> > "private" grid I've been using.
>
> More information needed! Could you please reproduce this problem and
> post the timing report (from the "Recent Uploads and Downloads" page)
> of an upload that is stalled? Also, quantify "stalled".
>
> I will try to re-point at the Voluneer Grid and grab the timing report next
week some time.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://tahoe-lafs.org/pipermail/tahoe-dev/attachments/20100306/9f707b11/attachment.html>