it is awesome to have a brain and know where the hands come from, isn't it? ;-)
THANK YOU Michael -- that looks like you solved our biggest problem and solved it right. Shame on me that I had started envisioning some complicated queue process -- locking imho is sufficient and due to simplicity of the implementation -- preferable. I have committed your patch upstream (sorry -- I haven't yet migrated it once and for all from SVN to GIT, so authorship is only in the comment) and uploaded fix to Debian. thanks again ;-) On Fri, 23 Sep 2011, Michael Saavedra wrote: > I created a patch for this bug. It is not very extensive. In fact it > changes just the bare minimum needed to prevent the problem. I don't > add a queue then have all the commands run by a single thread as > Yaroslav suggested, I just added a lock to executeCmd() in the > server/action.py file to prevent calling iptables or other shell > commands concurrently. > The patch can be found at: > https://sourceforge.net/tracker/?func=detail&aid=3413485&group_id=121032&atid=689046 > (my apologies if that link wraps to the next line) > Michael Saavedra -- =------------------------------------------------------------------= Keep in touch www.onerussian.com Yaroslav Halchenko www.ohloh.net/accounts/yarikoptic -- To UNSUBSCRIBE, email to [email protected] with a subject of "unsubscribe". Trouble? Contact [email protected]
