January 26, 2016 at 6:03 am
Brandie Tarvin (1/26/2016)
TomThomson (1/25/2016)
GilaMonster (1/25/2016)
Bandwidth, latency and data out of the country (for 1 and 3)I see the same. Plus in the EU it's probably illegal to host the data outside the EU if any of it is personally identifiable - we now have a formal decision from the court that it is illegal to host such data on servers in the USA or servers outside teh USA but under the control of American companies. Of course some new "safe harbor" farce will be cooked up and be treatedd as real until the court once again deems it fake.
Where I come from, safe harbor means places where parents can drop off infants and babies without being arrested for negligence or child abandonment.
What does it mean in your statement, Tom?
We have a similar law here in Michigan.
In this context, safe harbor is a set of laws the regulate how data about EU citizens is handled and stored. It's mean to protect privacy.
January 26, 2016 at 6:14 am
Brandie Tarvin (1/26/2016)
TomThomson (1/25/2016)
GilaMonster (1/25/2016)
Bandwidth, latency and data out of the country (for 1 and 3)I see the same. Plus in the EU it's probably illegal to host the data outside the EU if any of it is personally identifiable - we now have a formal decision from the court that it is illegal to host such data on servers in the USA or servers outside teh USA but under the control of American companies. Of course some new "safe harbor" farce will be cooked up and be treatedd as real until the court once again deems it fake.
Where I come from, safe harbor means places where parents can drop off infants and babies without being arrested for negligence or child abandonment.
What does it mean in your statement, Tom?
Change "infants and babies" and "child" for "data". :hehe:
January 26, 2016 at 7:37 am
I guess my main misunderstanding here is Tom's use of "'safe harbor' farce". I feel as if I'm missing a specific reference.
January 26, 2016 at 8:34 am
rodjkidd (1/26/2016)
In other news,Just ran one of Jason scripts!
So everyone back away from their keyboards for a minute....
🙂
Nah, it worked as expected...
All good here!
Rodders...
So all databases were dropped and backups deleted properly (secure wipe)?
Jason...AKA CirqueDeSQLeil
_______________________________________________
I have given a name to my pain...MCM SQL Server, MVP
SQL RNNR
Posting Performance Based Questions - Gail Shaw[/url]
Learn Extended Events
January 26, 2016 at 8:57 am
SQLRNNR (1/26/2016)
rodjkidd (1/26/2016)
In other news,Just ran one of Jason scripts!
So everyone back away from their keyboards for a minute....
🙂
Nah, it worked as expected...
All good here!
Rodders...
So all databases were dropped and backups deleted properly (secure wipe)?
Did you ever get the part working where it went out, found all backup tapes and irradiated them? It's in the same schema as the one where you find the cloud backups and issue a tactical nuke on the server.
January 26, 2016 at 9:02 am
Ed Wagner (1/26/2016)
SQLRNNR (1/26/2016)
rodjkidd (1/26/2016)
In other news,Just ran one of Jason scripts!
So everyone back away from their keyboards for a minute....
🙂
Nah, it worked as expected...
All good here!
Rodders...
So all databases were dropped and backups deleted properly (secure wipe)?
Did you ever get the part working where it went out, found all backup tapes and irradiated them? It's in the same schema as the one where you find the cloud backups and issue a tactical nuke on the server.
It did everything as advertised.
I am denying everything 😉
Rodders...
January 26, 2016 at 11:03 am
Is anyone else having a "dot your Ts and cross your Is" kind of a day?
January 26, 2016 at 11:53 am
Brandie Tarvin (1/26/2016)
Is anyone else having a "dot your Ts and cross your Is" kind of a day?
Gads yes...
:-/
January 26, 2016 at 1:40 pm
jasona.work (1/26/2016)
Brandie Tarvin (1/26/2016)
Is anyone else having a "dot your Ts and cross your Is" kind of a day?Gads yes...
:-/
Cross your eyes and roll your eyes definitely.
Jason...AKA CirqueDeSQLeil
_______________________________________________
I have given a name to my pain...MCM SQL Server, MVP
SQL RNNR
Posting Performance Based Questions - Gail Shaw[/url]
Learn Extended Events
January 26, 2016 at 2:13 pm
Grant Fritchey (1/25/2016)
Hugo Kornelis (1/25/2016)
Grant Fritchey (1/25/2016)
A few questions for the threadizens. Interested in this from everyone, but especially interested in what you guys who consult think.What blockers are you seeing for Azure?
What blockers are you seeing for a datawarehouse?
What blockers are you seeing for a datawarehouse on azure?
My current contract is actually working towards the cloud (for their OLTP system, they don't have a datawarehouse). So they see no blockers.
I am worried about bandwidth. Their current bandwidth is cr*p, but they plan to upgrade that before actually going cloud. I really hope that the upgrade will be sufficient.
Another major blocker is the existing system. They have several third-party systems that use SQL Server for data, and then some home-built systems that try to tie everything together - by using a variety of techiques (so far I have seen linked servers for querying and updating directly in the 3rd party database, I have seen replication, I have seen Agent jobs that forcefully sync full table data every X hours (or minutes!), and I have seen various "message queue"-based systems that capture changes from triggers on the 3rd party databases and feed that into BizTalk for further processing by other systems. Getting all those undocumented cross-database and cross-server dependencies documented and replaced by things that Azure actually supports will be a huge challenge.
That sounds more than a little challenging. So... if you could wish a tool into existence that helped you solve some of these problems for the migration, any idea what it might look like? I don't mean a magic tool, but some additional tooling not currently available that would make that process easier?
In my case, the majority of the cross-database contamination (I am not going to use any of the words interfacing or coupling here) is by the use of synonyms that then resolve to an object in either a different database on the same instance, or a database on another instance. I personally hate that with a vengeance. The cross-database and (especially!) the cross-server queries are terrible for performance. And the synonyms hide that; developers churn out code often not even aware how much data they are dragging between servers.
But in spite of my dislike, I must admit that for conversions those synonyms are nice. I could move a database to another server and just change the synonyms to ensure that the system continues to work. That same mechanism would facilitate a move to SQL Database - if cross-database references were supported. But they're not.
So the one tool that would really make a move to SQL Database feasible for my current customer would be support in SQL Database for cross-database queries. Getting all the security for those references set up properly, and then redirecting all the synonyms is something I can handle.
January 26, 2016 at 3:29 pm
Brandie Tarvin (1/26/2016)
I guess my main misunderstanding here is Tom's use of "'safe harbor' farce". I feel as if I'm missing a specific reference.
It was a specific agreement between the EU and the USA that contracts between European companies and American companies would be allowed to enforce EU standards of data protection for any of the EU companies' data that was stored and/or processed in the USA. It was a farce because everyone with a clue about security and teh way the US government works knew it was completely useless. The American govenment didn't do anything to enable those standards to be enforced in the USA and continued to legislate and regulate in such a manner as to ensure that the data would be fully available to every imaginable part of American officialdom without any protection at all. I've been warning people for years (ever since that "Safe Harbor" agreement was agreed) that the agreement was worthless because the USA had no intention of fulfilling its part, and that the EU courts would eventually rule that the agreement did not permit European personally identifiable data to be stored or processed in the USA or by any American company, and that "personally identifiable" was rather widely defined in EU law. The courts have now declared that the "safe harbour" agreement does not permit PI data in the hands of a company subject to EU law to be stored or processed in the USA; effectively they declared the agreement null and void.
Microsoft sems to be just about the only US company that actually understands this stuff, and it is trying hard to conform to European law for data that it processes in Ireland. I believe that it is still in dispute with some department of the US government becuse it's refusing to release any of this data to them, while the US government are claiming that even Microsoft's Irish subsidiary, operating in Ireland, is subject not to any EU law, in particular not to Irish law, in respect of p.i data it holds in Ireland, but only to US law. If Microsoft loses that battle with the US government it will probably lead (i) to the EU declaring that no US company can be trusted to obtain or hold any p.i. data in Europe, which will not be fun for the USA and EU economies, (ii) to a disastrous legal mess for Microsoft (the US government tends to get pretty vindictive when defied, but if \microsoft knowingly breaks Irish law the Irish could get pretty nasty too) and (iii) to a big diplomatic crisis between EU and USA.
It is now illegal for SWIFT to route European banking data through the USA, but I believe everyone's turning a blind eye to that until such time as SWIFT can make alternative arrangements.
The most scary thing about all this for the UK is that Obama seems to have convinced Cameron that the UK needs to have data protection law as useless as the USA has.
Tom
January 26, 2016 at 4:11 pm
Brandie Tarvin (1/26/2016)
Is anyone else having a "dot your Ts and cross your Is" kind of a day?
What ? :w00t: ?
Don't you have those every day ? 😉
Tom
January 26, 2016 at 4:12 pm
Brandie Tarvin (1/26/2016)
TomThomson (1/25/2016)
GilaMonster (1/25/2016)
Bandwidth, latency and data out of the country (for 1 and 3)I see the same. Plus in the EU it's probably illegal to host the data outside the EU if any of it is personally identifiable - we now have a formal decision from the court that it is illegal to host such data on servers in the USA or servers outside teh USA but under the control of American companies. Of course some new "safe harbor" farce will be cooked up and be treatedd as real until the court once again deems it fake.
Where I come from, safe harbor means places where parents can drop off infants and babies without being arrested for negligence or child abandonment.
What does it mean in your statement, Tom?
I think you are referring to "Safe Havens" Brandie. Safe Haven laws allow parents to drop off babies without getting in trouble.
-- Itzik Ben-Gan 2001
January 27, 2016 at 2:02 am
Hugo Kornelis (1/26/2016)
Grant Fritchey (1/25/2016)
Hugo Kornelis (1/25/2016)
Grant Fritchey (1/25/2016)
A few questions for the threadizens. Interested in this from everyone, but especially interested in what you guys who consult think.What blockers are you seeing for Azure?
What blockers are you seeing for a datawarehouse?
What blockers are you seeing for a datawarehouse on azure?
My current contract is actually working towards the cloud (for their OLTP system, they don't have a datawarehouse). So they see no blockers.
I am worried about bandwidth. Their current bandwidth is cr*p, but they plan to upgrade that before actually going cloud. I really hope that the upgrade will be sufficient.
Another major blocker is the existing system. They have several third-party systems that use SQL Server for data, and then some home-built systems that try to tie everything together - by using a variety of techiques (so far I have seen linked servers for querying and updating directly in the 3rd party database, I have seen replication, I have seen Agent jobs that forcefully sync full table data every X hours (or minutes!), and I have seen various "message queue"-based systems that capture changes from triggers on the 3rd party databases and feed that into BizTalk for further processing by other systems. Getting all those undocumented cross-database and cross-server dependencies documented and replaced by things that Azure actually supports will be a huge challenge.
That sounds more than a little challenging. So... if you could wish a tool into existence that helped you solve some of these problems for the migration, any idea what it might look like? I don't mean a magic tool, but some additional tooling not currently available that would make that process easier?
In my case, the majority of the cross-database contamination (I am not going to use any of the words interfacing or coupling here) is by the use of synonyms that then resolve to an object in either a different database on the same instance, or a database on another instance. I personally hate that with a vengeance. The cross-database and (especially!) the cross-server queries are terrible for performance. And the synonyms hide that; developers churn out code often not even aware how much data they are dragging between servers.
But in spite of my dislike, I must admit that for conversions those synonyms are nice. I could move a database to another server and just change the synonyms to ensure that the system continues to work. That same mechanism would facilitate a move to SQL Database - if cross-database references were supported. But they're not.
So the one tool that would really make a move to SQL Database feasible for my current customer would be support in SQL Database for cross-database queries. Getting all the security for those references set up properly, and then redirecting all the synonyms is something I can handle.
My jaw hit the floor when I read the last paragraph. No support for cross-database queries? Two databases on the same "instance" isn't allowed or the two databases can't reference each other using three-part naming? I've had a poke around in the release notes for elastic databases and a few other docs [/url] - but the over-liberal sprinkling of the keyword "remote" renders the documentation - at least for me - virtually unusable. I can't find an answer to the simple question "can two databases on the same instance communicate transparently as they can with terrestrial SQL Server". I keep landing on sharding - which, if it's meant to be a solution, appears to me to be massive overkill. Or maybe I'm particularly thick this morning.
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
January 27, 2016 at 2:19 am
ChrisM@Work (1/27/2016)
Two databases on the same "instance" isn't allowed or the two databases can't reference each other using three-part naming?
With Azure SQL DB, there's no such thing as 'instances'. You have a DB. Where it is, is none of your concern. It may be on the same 'server' as another DB today and elsewhere tomorrow. You have a SQL DB offered as a service, with all the server-level stuff handled for you.
If you want control over which DBs are on an instance, which instances are on a server, then you need to control the server, that's Azure VMs with SQL installed on them
Gail Shaw
Microsoft Certified Master: SQL Server, MVP, M.Sc (Comp Sci)
SQL In The Wild: Discussions on DB performance with occasional diversions into recoverability
Viewing 15 posts - 52,381 through 52,395 (of 66,738 total)
You must be logged in to reply to this topic. Login to reply