October 1, 2018 at 10:59 am
I keep waiting to see Joe C show up in the "Give Up on Natural Primary Keys" topic...
It'd be interesting to see what his take on it is, especially considering that it does sound like GDPR-type regulations could kill natural keys when you're dealing with people...
October 3, 2018 at 2:55 am
Ed Wagner - Monday, October 1, 2018 10:47 AMThanks to Mike for the tour of Pittsburgh, even though it was a bit crazy at times. π
Eirikur, it was great to meet you in person. I feel honored that you'd make such a long trip. It was a very good weekend, but certainly not long enough.Have a safe trip home, my friend.
Where are the beers???
For fast, accurate and documented assistance in answering your questions, please read this article.
Understanding and using APPLY, (I) and (II) Paul White
Hidden RBAR: Triangular Joins / The "Numbers" or "Tally" Table: What it is and how it replaces a loop Jeff Moden
October 3, 2018 at 3:08 am
Having a couple of pc's at home to play with is common for developers, but what about something geared up specifically for experimenting with SQL Server? There are a couple of things I'd like to spend some quality time on, the sort of time which cannot be hidden at work or masked as "learning" and for these I'd like to set up a reasonable home lab. A couple of things I'm considering as tasks:
- We've got cache-priming working a treat here, we can archive over a hundred million rows from a 20,000 million row table in quite a bit less than an hour without affecting other work on the server. I'd like to build a functional model and document this.
- Some example studies of separating out storage for data, log & tempdb. There's not enough of this around.
So yesterday I opened a "chat" with Dell, asking about a modest home server with 64GB RAM and three independently - writable storage areas. Ten minutes later they sent me a quote for something which wouldn't look out of place in a datacentre with a pricetag of over Β£12k!
How do you go about sourcing home kit at this level?
For fast, accurate and documented assistance in answering your questions, please read this article.
Understanding and using APPLY, (I) and (II) Paul White
Hidden RBAR: Triangular Joins / The "Numbers" or "Tally" Table: What it is and how it replaces a loop Jeff Moden
October 3, 2018 at 3:45 am
ChrisM@Work - Wednesday, October 3, 2018 3:08 AMHaving a couple of pc's at home to play with is common for developers, but what about something geared up specifically for experimenting with SQL Server? There are a couple of things I'd like to spend some quality time on, the sort of time which cannot be hidden at work or masked as "learning" and for these I'd like to set up a reasonable home lab. A couple of things I'm considering as tasks:
- We've got cache-priming working a treat here, we can archive over a hundred million rows from a 20,000 million row table in quite a bit less than an hour without affecting other work on the server. I'd like to build a functional model and document this.
- Some example studies of separating out storage for data, log & tempdb. There's not enough of this around.
So yesterday I opened a "chat" with Dell, asking about a modest home server with 64GB RAM and three independently - writable storage areas. Ten minutes later they sent me a quote for something which wouldn't look out of place in a datacentre with a pricetag of over £12k!
How do you go about sourcing home kit at this level?
For practicing storage, get some SD cards or thumbdrives. SQL will be slow, but to experiment, you don't need to be fast, you need to study differences. So you can get 3 class 10 SD cards which are fairly fast, for not much. Then get a few readers
Alternatively, you can get some cheap(ish) external SSDs and use those if you want more speed.
RAM is tough, but again, don't try to duplicate production. Scale down. Instead of 200mm rows, what about 2mm, or 200k rows, can you warm a cache there? You can still experiment, just scale down and extrapolate.
I'd suggest Azure, but that can get expensive, and it's easy to forget to turn off machines, or get caught up in faster SSD storage, which is $$$.
October 3, 2018 at 6:07 am
ChrisM@Work - Wednesday, October 3, 2018 2:55 AMWhere are the beers???
By that time, they were gone. Don't worry...they didn't drink all of them. π The picture was on Sunday when Mike took us on a tour of the winding roads through the hills of Pittsburgh.
October 3, 2018 at 6:10 am
October 3, 2018 at 7:29 am
jasona.work - Wednesday, October 3, 2018 6:10 AMAnother option would be to shop Craigslist (or your local equivalent,) for people / places selling their old equipment. Sometimes you can find a decent beast for a really decent price for setting up a home lab.
Example: I scored an older Dell r900 server, no drives or drive caddies (but both a Dell RAID card and space for 6 drives), with 4x quad core Xeons and 64GB of RAM for $200. Caddies added on about $20/ea, which I think popped in a couple old 2.5" SATA drives I had handy, then another $50 on the rack rails for the server. It's worked for a treat for my home lab, although it's a rather noisy beast when it's on, so it gets shut down when not in use.
It's definitely, though, more than enough for what I need in a home lab, but for the price, well...
π
I've put together various Frankenstein PC/Servers over the years built on anything I can beg, borrow, or steal.
For example, there is usually a 3 year cycle for servers. The 3 year old boxes are fully depreciated by the accountants, so the value is 0 to a company. They get donated, or sent to a recycler.
Look for companies that may be downsizing, or going out of business.
The local Goodwill stores take donations of PC's and sell them.
I actually use Azure for this testing, but you do need to be very vigilant or it will be a big bill!
Michael L John
If you assassinate a DBA, would you pull a trigger?
To properly post on a forum:
http://www.sqlservercentral.com/articles/61537/
October 3, 2018 at 7:36 am
Michael L John - Wednesday, October 3, 2018 7:29 AMI've put together various Frankenstein PC/Servers over the years built on anything I can beg, borrow, or steal.
For example, there is usually a 3 year cycle for servers. The 3 year old boxes are fully depreciated by the accountants, so the value is 0 to a company. They get donated, or sent to a recycler.
Look for companies that may be downsizing, or going out of business.I actually use Azure for this testing, but you do need to be very vigilant or it will be a big bill!
That works too, take an old desktop that you're replacing, even if it's just the guts, throw that in a new, cheap, case with some drives, and you've got a home-lab server.
October 3, 2018 at 7:41 am
jasona.work - Wednesday, October 3, 2018 7:36 AMMichael L John - Wednesday, October 3, 2018 7:29 AMI've put together various Frankenstein PC/Servers over the years built on anything I can beg, borrow, or steal.
For example, there is usually a 3 year cycle for servers. The 3 year old boxes are fully depreciated by the accountants, so the value is 0 to a company. They get donated, or sent to a recycler.
Look for companies that may be downsizing, or going out of business.I actually use Azure for this testing, but you do need to be very vigilant or it will be a big bill!
That works too, take an old desktop that you're replacing, even if it's just the guts, throw that in a new, cheap, case with some drives, and you've got a home-lab server.
That's probably the direction I'll take - the six x 500GB Barracuda drives in my drawer are influencing the decision.
For fast, accurate and documented assistance in answering your questions, please read this article.
Understanding and using APPLY, (I) and (II) Paul White
Hidden RBAR: Triangular Joins / The "Numbers" or "Tally" Table: What it is and how it replaces a loop Jeff Moden
October 3, 2018 at 8:15 am
ChrisM@Work - Wednesday, October 3, 2018 2:55 AMWhere are the beers???
What... you missed the side profile of my gut in the picture? π
--Jeff Moden
Change is inevitable... Change for the better is not.
October 3, 2018 at 8:20 am
Jeff Moden - Wednesday, October 3, 2018 8:15 AMWhat... you missed the side profile of my gut in the picture? π
A friend's 12-year old daughter calls his steadily-increasing gut his "food baby" π
"Beer baby" works too.
For fast, accurate and documented assistance in answering your questions, please read this article.
Understanding and using APPLY, (I) and (II) Paul White
Hidden RBAR: Triangular Joins / The "Numbers" or "Tally" Table: What it is and how it replaces a loop Jeff Moden
October 3, 2018 at 9:59 am
Steve Jones - SSC Editor - Wednesday, October 3, 2018 3:45 AMChrisM@Work - Wednesday, October 3, 2018 3:08 AMHaving a couple of pc's at home to play with is common for developers, but what about something geared up specifically for experimenting with SQL Server? There are a couple of things I'd like to spend some quality time on, the sort of time which cannot be hidden at work or masked as "learning" and for these I'd like to set up a reasonable home lab. A couple of things I'm considering as tasks:
- We've got cache-priming working a treat here, we can archive over a hundred million rows from a 20,000 million row table in quite a bit less than an hour without affecting other work on the server. I'd like to build a functional model and document this.
- Some example studies of separating out storage for data, log & tempdb. There's not enough of this around.
So yesterday I opened a "chat" with Dell, asking about a modest home server with 64GB RAM and three independently - writable storage areas. Ten minutes later they sent me a quote for something which wouldn't look out of place in a datacentre with a pricetag of over £12k!
How do you go about sourcing home kit at this level?For practicing storage, get some SD cards or thumbdrives. SQL will be slow, but to experiment, you don't need to be fast, you need to study differences. So you can get 3 class 10 SD cards which are fairly fast, for not much. Then get a few readers
Alternatively, you can get some cheap(ish) external SSDs and use those if you want more speed.
RAM is tough, but again, don't try to duplicate production. Scale down. Instead of 200mm rows, what about 2mm, or 200k rows, can you warm a cache there? You can still experiment, just scale down and extrapolate.
I'd suggest Azure, but that can get expensive, and it's easy to forget to turn off machines, or get caught up in faster SSD storage, which is $$$.
Thanks Steve. I'm planning to implement the 3 x RAID option but a comparison with three thumb drives will be fun to do.
"Your io is terrible because everything's on the same disk - you'd be better off with three thumb drives".
For fast, accurate and documented assistance in answering your questions, please read this article.
Understanding and using APPLY, (I) and (II) Paul White
Hidden RBAR: Triangular Joins / The "Numbers" or "Tally" Table: What it is and how it replaces a loop Jeff Moden
October 3, 2018 at 10:07 am
ChrisM@Work - Wednesday, October 3, 2018 9:59 AMThanks Steve. I'm planning to implement the 3 x RAID option but a comparison with three thumb drives will be fun to do.
"Your io is terrible because everything's on the same disk - you'd be better off with three thumb drives".
I've got some spare stuff you can have Chris, shall we talk through what you need over a beerπ
π
October 3, 2018 at 10:21 am
FYI: Totally non-DBA job, but if anyone knows someone looking for a job in Jacksonville... https://careers.allstate.com/job/Jacksonville-Workforce-Management-and-Strategic-Analytics-Flor/509833200/?locale=en_US
October 3, 2018 at 12:22 pm
I can't decide if the posted replies here are just trolling the OP or what: https://www.sqlservercentral.com/Forums/1998406/Automatically-avoid-error-inserting-into-identity-field
Although I think I'm leaning towards trolling...
Viewing 15 posts - 62,371 through 62,385 (of 66,712 total)
You must be logged in to reply to this topic. Login to reply