A secondary site, regardless of scale was a bad design choice
From: [email protected] [mailto:[email protected]] On Behalf Of Santos Martinez Sent: 05 October 2015 20:09 To: [email protected] Subject: RE: [mssms] SCCM Redundancy Best Practices Agree with Jason on this, I will suggest to review carefully the following documentation. https://technet.microsoft.com/en-us/library/gg682077.aspx#BKMK_SiteAndRoleScale Currently, we support up to 10,000 clients in the secondary site with “SP2”. Best Regards, Santos From: [email protected] <mailto:[email protected]> [mailto:[email protected]] On Behalf Of Jason Sandys Sent: Monday, October 5, 2015 2:57 PM To: [email protected] <mailto:[email protected]> Subject: RE: [mssms] SCCM Redundancy Best Practices Using what technology though? Simply having an additional primary site is meaningless as far as HA goes – there is no way to move your clients across so it’s completely useless. A secondary site is even worse since everything about a secondary site is a single point of failure. Square pegs, round holes. J From: [email protected] <mailto:[email protected]> [mailto:[email protected]] On Behalf Of Jason Wallace Sent: Monday, October 5, 2015 1:51 PM To: [email protected] <mailto:[email protected]> Subject: Re: [mssms] SCCM Redundancy Best Practices I have seen a couple of designs where the thinking was that having a PRI in each Datacenter would allow you to continue to run the environment in the event of a complete Datacenter failure. I have even seen one where they then (within the Datacenter) scaled out a new secondary site for every 25,000 clients. Not my designs I hasten to add On 5 Oct 2015, at 18:40, Jason Sandys <[email protected] <mailto:[email protected]> > wrote: What’s an offsite primary server? Are you talking about a cold-stand-by server that you would perform a restore on? J From: [email protected] <mailto:[email protected]> [mailto:[email protected]] On Behalf Of Charles Hiland Sent: Monday, October 5, 2015 12:32 PM To: [email protected] <mailto:[email protected]> Subject: RE: [mssms] SCCM Redundancy Best Practices Thanks, Jason and others for the input, and Garth for the laugh J, and please forgive my semantics – I didn’t mean secondary sites (to me if they’re not primary, they’re secondary. Apologies!). I think from what I’ve heard from feedback is that our best option moving forward may be to have an off-site primary server. We currently have an on-site primary, and want to have a setup where an office being out of commission (e.g. Hurricane Sandy) doesn’t impact all of our smaller, remote offices. Thank you for the feedback! From: [email protected] <mailto:[email protected]> [mailto:[email protected]] On Behalf Of Jason Sandys Sent: Monday, October 05, 2015 1:16 PM To: [email protected] <mailto:[email protected]> Subject: RE: [mssms] SCCM Redundancy Best Practices +150,000 From: [email protected] <mailto:[email protected]> [mailto:[email protected]] On Behalf Of Ryan Sent: Monday, October 5, 2015 12:12 PM To: [email protected] <mailto:[email protected]> Subject: Re: [mssms] SCCM Redundancy Best Practices On Mon, Oct 5, 2015 at 12:01 PM, Charles Hiland <[email protected] <mailto:[email protected]> > wrote: The end result is redundancy in our primary server and SQL database. From: [email protected] <mailto:[email protected]> [mailto:[email protected] <mailto:[email protected]> ] On Behalf Of Jason Wallace Sent: Monday, October 05, 2015 12:47 PM To: [email protected] <mailto:[email protected]> Subject: Re: [mssms] SCCM Redundancy Best Practices Nooooooo! How many clients do you have? In CM16 you can have 200,000 clients. What do you expect to see from your new design? On 5 Oct 2015, at 17:41, Charles Hiland <[email protected] <mailto:[email protected]> > wrote: We will be upgrading to SCCM 2016 next year, and I’m trying to figure out an ideal redundancy setup. At the moment we have one primary and 7 secondary sites (the secondary sites are just DPs). Does anyone have any ideas for best practice when it comes to redundancy? Currently we are thinking an off-site CAS with two primary servers. Do any common-place rules stand out that I am overlooking? Thanks in advance, Charlie
