Morning

“The scariest thing about COVID-19 is that it looks like a JIRA ticket”

Oh so true

Haha

Typical programmers trying to build everything themselves

Well you'd need to code such a thing yourself because no sensible database would provide built-in features for that.

@PaulWhite9 did you understand what they were asking? I would have closed as unclear.

@ypercubeᵀᴹ No, but I don't know anything about mongodb so didn't 'vote'.

ah, I didn't even read the tags ...

@MasterDatabase We should post an ad for topanswers.xyz

Quick and dirty ad ^^ :D

the community that trolls together stays together

Who says that's trolling?

> The goal is for future visitors to find out about the stuff your community deems important. This also serves as a way to promote information and resources that are relevant to your own community's interests, both for those already in the community and those yet to join.

Seems relevant

i think the definition of "community" in the official sense is probably a different entity than the community that you and i are a part of - that is the one that treats the membrane between TA & SE as semi-permeable

@Colin'tHart That's hilarious - that's small enough to just replace every night and do some kind of simple swap. No need for all those shenanigans.

topanswers.xyz/…

☝ also, that

tbh i had the same couple thoughts, but i figured Jack will be aware the community adds and make his decision in his own time. he needs to decide whether he wants to re-up db<>fiddle in any case

@PeterVandivier The community in this scenario is basically whoever bothers to vote

and the company that has veto power probs 🤷‍♂️

all of which is to say, i'm absolutely gonna upvote any ad jack decides to post

just don't have 100% confidence in a TA ad specifically being allowed to stick around - especially if it becomes a noticeably anti-authoritarian talking point

@TomV-trytopanswers.xyz vmware 6.7

No NUMA issues. 8 vCores on same NUMA node.

256 GB RAM

@Johnakahot2use for the VM or for the physical box?

Physical box

ESX host has 768 GB RAM

+---------------------+--------+------------+----------+-----------+------------+-----------+----------+----------+
|      WaitType       | Wait_S | Resource_S | Signal_S | WaitCount | Percentage | AvgWait_S | AvgRes_S | AvgSig_S |
+---------------------+--------+------------+----------+-----------+------------+-----------+----------+----------+
| LCK_M_U             |  39.83 |      39.81 |     0.02 |        29 |      10.05 |    1.3734 |   1.3728 |   0.0006 |
| CXCONSUMER          |  22.31 |      22.08 |     0.23 |       705 |       5.63 |    0.0316 |   0.0313 |   0.0003 |

Waits occurring during a 5 minute period when performance diminishes.

Sorry that should read: 256 GB RAM for the virtual server. The ESX host has a total of 768 GB RAM.

And the virtual server is currently running on its own ESX host.

@Johnakahot2use Do you have a baseline for when it's running ok?

@Johnakahot2use Should be able to allocate all vm ram on the numa node the vcpu's are allocated on then

@TomV-trytopanswers.xyz yep

"\\tsclient\s\tools\Sysinternals Suite\Coreinfo.exe"


Maximum implemented CPUID leaves: 0000000D (Basic), 80000008 (Extended).

Logical to Physical Processor Map:
*-------  Physical Processor 0
-*------  Physical Processor 1
--*-----  Physical Processor 2
---*----  Physical Processor 3
----*---  Physical Processor 4
-----*--  Physical Processor 5
------*-  Physical Processor 6
-------*  Physical Processor 7

Logical Processor to Socket Map:
*-------  Socket 0
-*------  Socket 1
--*-----  Socket 2

and ...

vmdumper -l | cut -d \/ -f 2-5 | while read path; do egrep -oi "DICT.*(displayname.*|numa.*|cores.*|vcpu.*|memsize.*|affinity.*)= .*|numa:.*|numaHost:.*" "/$path/vmware.log"; echo -e; done


DICT                  numvcpus = "8"
DICT                   memSize = "262144"
DICT               displayName = "GMDMSSQLP07"
DICT numa.autosize.vcpu.maxPerVirtualNode = "8"
DICT      numa.autosize.cookie = "80001"
numaHost: NUMA config: consolidation= 1 preferHT= 0
numaHost: 8 VCPUs 1 VPDs 1 PPDs
numaHost: VCPU 0 VPD 0 PPD 0

@TomV-trytopanswers.xyz Sadly not.

We have values for when it seems to be performing ok.

+---------------------+--------+------------+----------+-----------+------------+-----------+----------+----------+
|      WaitType       | Wait_S | Resource_S | Signal_S | WaitCount | Percentage | AvgWait_S | AvgRes_S | AvgSig_S |
+---------------------+--------+------------+----------+-----------+------------+-----------+----------+----------+
| CXCONSUMER          | 553.24 |     545.77 |     7.47 |      4230 |      57.68 |    0.1308 |    0.129 |   0.0018 |
| CXPACKET            |  193.4 |     190.58 |     2.82 |      2089 |      20.16 |    0.0926 |   0.0912 |   0.0013 |

and...

+---------------------+--------+------------+----------+-----------+------------+-----------+----------+----------+
|      WaitType       | Wait_S | Resource_S | Signal_S | WaitCount | Percentage | AvgWait_S | AvgRes_S | AvgSig_S |
+---------------------+--------+------------+----------+-----------+------------+-----------+----------+----------+
| CXCONSUMER          |  81.94 |      63.09 |    18.85 |     14131 |      44.41 |    0.0058 |   0.0045 |   0.0013 |
| SOS_SCHEDULER_YIELD |  45.88 |       0.03 |    45.84 |     77852 |      24.86 |    0.0006 |        0 |   0.0006 |

No blocking ^^

Which there was in the first results

RCSI

+---------------------+--------+------------+----------+-----------+------------+-----------+----------+----------+
|      WaitType       | Wait_S | Resource_S | Signal_S | WaitCount | Percentage | AvgWait_S | AvgRes_S | AvgSig_S |
+---------------------+--------+------------+----------+-----------+------------+-----------+----------+----------+
| LCK_M_U             |  39.83 |      39.81 |     0.02 |        29 |      10.05 |    1.3734 |   1.3728 |   0.0006 |

Not terribly much, but still

That's when the new tax form are imported into the database.

Can RO redact things here?

@Johnakahot2use They can move things to trash, but not delete.

From the Coronavirus desk, seen on the interwebs today ...

Man, I like both of those, can't decide, will alternate methods.

@PaulWhite9 RO?

@PeterVandivier room owner

Ty

@billinkc i head you can fortify your immune system by licking hand rails

Man, Stack Overflow is a mess. Search on almost anything and you'll get a question with a few good answers and then a ton of answers that add absolutely nothing, just noise.

@MasterDatabase you need to go away, Master.

someone pull the SAN out from underneath that sucker

A really well asked question worth an upvote.

Even if questions about "best way" are normally closed as opinion-based, this one shows thought.

maybe it should still be closed, but i gave it an upvote for effort and an interesting idea & well constructed post

Scenario

I have SQL Server Dev 2017 installed

I want to install SQL Server Dev 2019

Should I uninstall 2017 or will the upgrade "remove" the old version for me?

@jcolebrand in-place is fine

I concur

i just did that.

thanks, thought so

to be fair, i haven't scoured my c:\ drive to see what all's been removed, but i used to select @@version and get 2017, and i don't anymore

I always just make instances, so I have .\SQL2012, .\SQL2014, .\SQL2016 etc.