Actually we just ran Resource Monitor manually. We never had any kind of automated system in place specifically for disk queues. Usually that caused enough other problems that we had plenty of alerts
----- Dwayne Allen [email protected] (479) 310-0027 On Wed, Jan 28, 2015 at 1:29 PM, Sean Pomeroy <[email protected]> wrote: > That would be correct, SQL co-located. > > We do have an issue that whomever configured the system made one big VHD > and diced it up into partitions for content, SQL and logs. > Makes it a pain to expand partitions. It was dynamic too! > I created separate VHDs to split up the partitions, just haven't had a big > enough maintenance window to move everything yet. > Maybe I should push for SSDs at the same time... > > Did you just use PerfMon to monitor disk queue? > > On Wed Jan 28 2015 at 2:20:16 PM Dwayne Allen <[email protected]> > wrote: > >> We did have some issues with our disk queues until we moved to SSDs so >> that is something you will want to keep an eye. I'm assuming you are >> co-locating SQL? >> >> ----- >> Dwayne Allen >> [email protected] >> (479) 310-0027 >> >> On Wed, Jan 28, 2015 at 10:08 AM, Sean Pomeroy <[email protected]> >> wrote: >> >>> @Nick, I currently have it configured with a SUP in each region. >>> I also have two IBCM SUPs configured for Americas and EMEA. >>> But we only have roughly ~1,500 clients internet only at any given time. >>> >>> @Dwayne, >>> Our Primary is a VM on a decently equipped box. >>> It's currently assigned 4 CPUs and 64GB of ram. >>> And we have plenty of resources free to assign more. >>> The only thing we don't have is dedicated Disks/SSDs. :( >>> >>> On Wed Jan 28 2015 at 10:37:38 AM Dwayne Allen <[email protected]> >>> wrote: >>> >> I think you maybe ok. One of our primary SUPs has over 20,000 clients >>>> and all of those roles outside of the RSP. It is a pretty beefy box >>>> though. >>>> >>>> ----- >>>> Dwayne Allen >>>> [email protected] >>>> (479) 310-0027 >>>> >>>> On Wed, Jan 28, 2015 at 9:07 AM, Sean Pomeroy <[email protected]> >>>> wrote: >>>> >>>>> We are in the process of planning and testing patch deployment via >>>>> SCCM. >>>>> >>>>> A high level overview of our infrastructure is as follows: >>>>> Primary in EMEA (MP, SUP, ACWP, DP, RSP, AISP) >>>>> Secondary in Americas (MP, SUP, DP) >>>>> Secondary in APAC (MP, SUP, DP) >>>>> >>>>> ~17,000 Clients (35% EMEA, 40% Americas, 20% APAC) >>>>> ~2,000 Servers (45% EMEA, 45% Americas, 10% APAC) >>>>> >>>>> At first we will only be patching clients. But servers will follow. >>>>> >>>>> Right now we are only testing patching on about ~500 PCs and all is >>>>> going smoothly. >>>>> >>>>> Once we move all clients to patching via SCCM, I am concerned our >>>>> Primary is going to be over loaded. >>>>> >>>>> I know MS documentation suggests SUP can handle 25,000 clients when it >>>>> co-exists with another role... But our Primary essentially has all roles. >>>>> >>>>> Should I push to stand up a separate server as a SUP in each region? >>>>> Or at least in place of the SUP on our primary? >>>>> >>>>> Thanks, >>>>> Sean >>>>> >>>>> >>>> >
