puppet, chef or ansible are more geared towards "prepare me a new node to add to the cluster" rather than "update the application files on every node of my cluster" but feel free to engage. My point was "start small and simple, then reiterate". chef ansible and puppet are more likely the 1000th reiteration than the 1st.
On Wednesday, August 12, 2015 at 5:50:10 PM UTC+2, Michael M wrote: > > That gives me a lot to go on. Thank-you Niphlod. > > I just got off the vanpool this morning and found out one of my co-workers > runs some of our corporate infrastructure. He suggested I look at Puppet, > Chef, or Ansible for cluster management. So looks like I have some > learning to do. :) > > Wish me luck. :) > > On Wednesday, August 12, 2015 at 12:03:33 AM UTC-7, Niphlod wrote: >> >> the update process on several frontend servers can be split into 2 modes: >> you can have an "incremental" update, in which the newer version of the >> application doesn't break the old one, and the "breaking" update, in which >> the newer version of the application breaks the old one. >> >> in the first case, you can log into each one of the servers and update >> the app. >> in the second case, you need to log into each server, put the application >> in "maintenance mode", update each server, and then revoke the "maintenance >> mode" on each. >> >> there are SEVERAL ways to do it, some more elegant and intricated, some >> more simple and maintainable. >> >> Start with basics: prepare a makefile or a procedure that updates a >> single node, then execute it in a "foreach server" loop. It can be bash, >> python, go, whatever you're more comfortable with. Draw the procedure on a >> sheet of paper first (yes, it gets the deal done to unravel the necessary >> steps beforehand) and then implement each piece. I use fabric to handle the >> automation part wherever I can. Bash makefiles or powershell scripts where >> I can't. >> >> On Wednesday, August 12, 2015 at 7:49:13 AM UTC+2, Michael M wrote: >>> >>> Hello All, >>> >>> I was looking at Load Balancing due to my Self Service Web Portal is I/O >>> and Networking intensive. >>> >>> I found >>> http://www.web2pyslices.com/slice/show/1360/high-availability-cluster-with-pound >>> >>> in it there is a this "We recommend using >>> <http://www.web2pyslices.com/slice/show/1360/29>this install procedure >>> <http://www.web2pyslices.com/slice/show/1360/29>" Which looks like a >>> dead link. >>> >>> So if anyone knows the current slice that is referring to would be a >>> great help. >>> >>> The main question is: Say I have 10 servers and I make a change to a >>> module and a view of an application. How do I update to all servers >>> without having to log into each and upload a Tar? >>> >>> Or what is the best practices? >>> >>> I am new to web-servers so please be verbose so I can learn as much as >>> possible. Thank-you. >>> >> -- Resources: - http://web2py.com - http://web2py.com/book (Documentation) - http://github.com/web2py/web2py (Source code) - https://code.google.com/p/web2py/issues/list (Report Issues) --- You received this message because you are subscribed to the Google Groups "web2py-users" group. To unsubscribe from this group and stop receiving emails from it, send an email to web2py+unsubscr...@googlegroups.com. For more options, visit https://groups.google.com/d/optout.
