SQL 2012 :: 500GB Data Warehouse - How To Split It Into File Groups
Aug 21, 2014
OK, so I have:
- 500 GB DW
- 5 GB in smaller DBs
- 220 GB TempDB
- 350 GB in Log files.
My machine is Fujitsu Primergy 64 cores (with HT) and 192 GB RAM.
I have several IO locations:
- 540 GB in-server HDD 15k RAID10
- 1 TB HDD 15k RAID10 on SAN (separete controller)
- 2 TB HDD 15k RAID10 on SAN (same controlller as below)
- 800GB SSD RAID10 on SAN (same controller as above)
Data warehouse has 2 fact tables that are absolutely crucial and quite large.
Now i want to organize DB into several Filegroups and put them on different drives. Filegroups I'm thinking of:
- FILEGROUP1: for 1st crucial Fact Table
- FILEGROUP2: for 2nd crucial Fact Table
- FILEGROUP3: for tempDB
- FILEGROUP4: for dimensions data
- FILEGROUP5: for the rest of facts data
- FILEGROUP6: for dimensions indexes
- FILEGROUP7: for the rest of facts indexes
- FILEGROUP8: for 1 log file of one smaller DB (its in full-recovery and its quite large)
- FILEGROUP9: for the rest of log files
- FILEGROUP10: others
How should I organize them across available drives? I was thinking about sth like:
I know that having multiple filegroups on the same drive is pointless regarding performance, but in future i could actually add some more drives, so i want to separate them now.
Also - how much files per filegroups should i create? Considering 1 or 2. Except TempDB where I am going for 4.
I have one of our production Accounting Databases starting from 2 GBnow grown into a 20 GB Database over the period of a few years...I have been getting timeouts when transactions are trying to updatedifferent tables in the database.. Most of the error I get are I/Orequests to the data file (Data file of the production dbAccounting_Data.MDF).I would like to implement the following to this Accounting database.I need to split the Data file into multiple files by placing some ofthe tables in different file groups. I have the server upgraded to beable to have different drives in different channels. I can place thesedata and log files in different drives so it will be less I/Oconflicts..I would like to have the following file groups..FileGroup 1 - which will have all database definitions (DDL).FileGroup 2 - I will have the AR Module tables under here..FileGroup 3 - I will have the GL module tables under here..FileGroup 4 - I will have the rest of the tables under hereFileGroup 5 - I will like to place the indexes under here....Also where will the associated transaction files go?I would like to get some help doing this. Is there any articles / helpavailable that I can refer to. Any suggestions / corrections/criticisms to what I have mentioned above is much appreciated...!Thanks in advance....
I am looking for the easiest way of rebalancing data across multiple files.
Instead of creating a secondary filegroup and then dropping and recreating all indexes in the database which is going to take ages (we have a lot of tables and indexes), I am trying to just add more files to the primary file group and then rebalance data evenly between these.
I guessed that adding the new files to the primary file group and then rebuilding all indexes on a table should redistribute the table over these multiple file groups evenly. This is not the case though. It does rebalance data a bit but I still end up with the majority on the first file that existed.
I have attached the script I am running, maybe it is something in the create database/file statements that is the issue.
Basically what I am seeing is to start off with the table is 160MB. I then add the file groups and rebuild all indexes on the table. The first file is then about 100MB and each of the three other files are about 20MB. I would expect them all to be the same size.
I need a recommendation on a data modeling tool that can be used with a data warehouse. My warehouse is running SQL 2012.
Here is my challenge: Most of the tables in the warehouse do not have primary keys and none of the tables have foreign keys on them. However, there are indexes and unique keys/indexes on the tables. I am looking for a tool that I can create virtual relationships on how the data is related, so it is visually easier for the ETL developers to write the code.
I have looked at both ER/Studio 11 and ERwin 9.6. Neither of them do it exactly the way I want it too. However, ER/Studio is pretty close.
We have a large Datawarehouse and the size is 50TB.. The tables are placed in filegroups based on the schema like fact, dimensions, raw data each sit on seperate filegroups. I am thinking will it make sense to seperate the large facts which are having billions of rows so that they reside on filegroups on their own..
When the database is configured for mirroring and you want to do partitioning on that database, How can we do? Is this similar process or any variation there while adding file groups and files? The partition will reflect in the mirroring database also?
I have a two node SQL 2012 AlwaysOn HADR cluster (v11.0.3412) with 4 availability groups configured. The AG groups are set to synchronous mode and the secondary is not readable (we do not want the synchronous replica readable so we do not risk any reads causing contention so we maintain fast performance).
On the secondary we are getting a persistent failure with the Data Collector job called Collection_Set_3_Upload. The failure occurs within the second job step. That job step is executing the following command:
dcexec -u -s 3 -i "$(ESCAPE_DQUOTE(MACH))$(ESCAPE_DQUOTE(INST))" The error message is as follows:
Log Job History (collection_set_3_upload) Step ID 2 Server CLUSTERNODE2 Job Name collection_set_3_upload Step Name collection_set_3_upload_upload Duration 00:00:07
[Code] ....
I know I can prevent this error message by enabling readable secondaries, but we do not want this.
I have tried stopping the data collection jobs and purging the cache directory but to no avail. It will succeed the first time then persistently fail again with the same message every time after that.
In addition, if I set the one failing AG group to readable secondary the job succeeds. So that means that 3/4 work fine, only this one is having an issue.
1 ,AU-Australia 1,MM-Myanmar 1,NZ-New Zealand 1,PG-Paua New Guinea 1,PH-Phlippines
Note: we are getting source data from sqlserver tables.
I googled and found below way but did't get the output as required
SELECT A.id, a.country, Split.a.value('.', 'VARCHAR(500)') AS String FROM (SELECT id, country , CAST ('<M>' + REPLACE(country, ' ', '</M><M>') + '</M>' AS XML) AS String FROM #t3) AS A CROSS APPLY String.nodes ('/M') AS Split(a);
I have a new SQL 2005 (SP2) Reporting Services server to which I've just upgraded and deployed some SSRS 2000 reports.
I have a subreport that contains a matrix with two groups. The report data seems to be inexplicably repeating the data for the first row in the group for all rows in the group. Example:
ID1 ID2 DisplayData
1 1 A
1 2 B
1 3 C
2 1 A
2 2 B
2 3 C
Parent group is on ID1, child group is on ID2, report would show:
1 1 A
2 A
3 A
2 1 A
2 A
3 A
Is this a matrix bug in 2005 SP2, or do I need to do something differently? I can no longer pull a comparison version from an SSRS 2000 server to verify, but I believe it was working as expected before...
I have a delimited text file with 650+ columns. The sum of the column lengths of a single row, if fully populated, exceeds 30K bytes. The "killer" fields lengthwise are the "Description" fields. If they were removed from the input file, the remainig columns would occupy about 5000 bytes, which is within SQL max row length.Â
Can SSIS be used to created these two tables? (one without  description fields, the other with those field but arranged vertically in the table rows).
The fundamental issue is I can not import a single file row into a sql table because that row length could exceed the max byte count for a row.
I have a requirement to implement CDC for 50+ tables to implement incremental data changes warehouse/reporting rather than exporting the whole table data. The largest table is having more than half a billion records.
The warehouse use a daily copy of OLTP db (daily DB refresh). How can I accomplish this. Is there a downside in implementing CDC just for the sake of taking incremental changes on the tables?
Is there any performance impact if we enable CDC on OLTP db?
Can we make use of the CDC tables on the environment we do daily db refresh so that the queries don't hit OLTP database?
What is the best way to implement CDC to take incremental changes for reporting.
hi I am new at MSSQL 2000 DBA thing. and trying to learn more about analysis service/data warehouse/data mining. so is any expert out there can Recommend some good books or web link article to read? Thanks
I am Crystal Reports Developer and I am new in SSIS environment. I have started to read Professional SQL Server 2005 IS book. I am really confused by many tasks to choose.
I need to develop reports from data warehouse. But before I have to send the data from operational database (SQL Server 2000) to warehouse (SQL Server 2005) monthly - I have a script for retrieving the data. For my package, I chose Data Flow Task, Execute SQL Task, and OLE DB Destination, and it does not work.
Please help me if I can look similar packages performing? Thank you!!
hi, Can anyone list the type of error that will make @@error =1 I created a procedure to update a table based on a cutomer id, Id 7 doest not exist in table A, and I am suppose to have Not valid id, but in this case nothing happen I always get table a updated thanks Ali
begin tran update table a set title =' manager ' where id =7
if (@error <> 0) begin rollback tran print 'not valid id ' return end ELSE begin commit tran print 'table A updated' end
I am working on to create a data warehouse. I have made a database which will be the data warehouse and will consist of dimension and fact tables. I know that other than dimension and fact table a data warehouse should also consist of a meta data, now my question is what should be the structure of metadata and all the information it should have?
I am going to use Microsoft SQL Server to develop my data warehouse, but one thing makes me confused. Since Analysis Service can create a Star schema database, do I have to pre-set up a Star schema database for ETLed data? Basically, I am wondering what's the relationship between an ETLed database and the one created through the Analysis Services.
Can any one give me an explanation from the implementation perspective?
I am trying to restore my data warehouse from a January 2008 backup under a new name to recover a table that I accidentally deleted. It is taking a long time for the restore to get done. Here is the command I am running as sa in QA
---
RESTORE DATABASE Warehouse_new FROM DISK = 'H:MSSQLDataMSSQLBACKUPDBBackupsWarehouseWa rehouse_db_200801050600.BAK' WITH MOVE 'Warehouse_Data' TO 'G:MSSQLDataMSSQLDataWarehouse_New_Data.MDF', MOVE 'Warehouse_Log' TO 'H:MSSQLDataMSSQLLogsWarehouse_New_Log.ldf'
----
There Warehouse_New_Data.MDF is 375 GB and the log is 12 GB.
There is still 169 GB of free space on the drive I am restoring to after the presence of Warehouse_Data.MDF and Warehouse_New_Data.MDF (each 375 GB).
Its been 4.5 hrs and the restore is still running. Backups take about 3.5 hrs to complete. Can I do any checks on the restore to see what point it is at? I stopped the restore using EM earlier after it took 8 hours and still no progress.
Hello..I was wondering if anyone out there could tell me how they deal withNULL values in a data warehouse? I am looking to implement a warehousein SQL 2005 and have some fields which will have NULL values and Iwould like some further ideas on how to deal with them. At my last jobin dealing with Oracle we were just going to leave the fields NULL, butin SQL how would you best recommend cleaning the data? I greatlyappreicate your help and look forward to your reponses.Thank you
I€™m making warehouse for our HMIS (healthcare management information system)by using SSIS. I€™m facing some problems now, could you please help me to solve my problem.
Brief idea about my Warehouse: Source: oracle 9i Destination: Sql server 2005 ETL tool: SSIS
Problems:
How to refresh or load the current data to data warehouse.(now i'm using truncate sql task for deleting old/entire data for each packages, i really dont want to use in the production) . For example: The patient admissions data is adding everyday so i want to load the current data into my warehouse. Could you pls suggest me good solution for this?
Refresh Cycle timings: is there any task available in SSIS?
current status:
First Time load completed, i set one Execute Sql statement ctrl flow task for Truncate the existing loaded data in the sql server 2005. and then again i process one data flow task for loading the data from oracle to sql server.
Is SSIS a tool for extracting realtime data from staging to data warehouse? Realtime in my case can be loading every 15 minutes but no more than 30 minutes. I've a data warehouse which data refresh once a day and it worked fine. The data that I extract into the warehouse is from a Staging database which is realtime replication of multiple production databases. Once a day, I've to have replication pauses on staging for a couple hours to refresh the data warehouse. That's the only way so SSIS can pull the data correctly; if I've replication on while SSIS pull data, it will always copy less rows than it supposed to.
I cannot afford to have replication pauses every 15 minutes just so I can refresh data warehouse. Does anyone every have this problem? or any best practice how to do this?
Does any body have the experience to execute data warehouse backup/recover? What I want to know is how to backup/recover database in data warehouse and cubes.
At my office, we've been slowly working on putting together a data warehouse.
We're a financial services company and one of the services that we offer is debt collection. As far as reports go, our clients are interested in knowing how much money we collect over time. In particular, they want to know how many payments we've gotten 5, 10, and 15 months (and so on) after we receive a case. (Obviously, the 5-month payments are also included in the 10 and 15-month calculations).
When I wrote this report using our transactional database, I was completely new to SQL and the ever-resourceful Patron Saint took pity on me, so you can see a good description of the details at http://www.sqlteam.com/forums/topic.asp?TOPIC_ID=78510
Now that I'm no longer a total newbie at SQL, and having been through a relatively extensive seminar on data warehousing, I've been entrusted with researching certain aspects of data warehouse development (rest easy, though, folks - the real DWH work is not being done by the very inexperienced me, but by an actual professional :) ).
My question:
how would you model this kind of "relative time" in a data warehouse? How would you display the 5-month, 10-month, and 15-month payments in a DWH? I can't really imagine that the kinds of joins necessary to do this in a transactional database would be desirable in a data warehouse.
We have the following:
1.) FACT_Payment: A fact table showing each each payment to the most detailed granularity. One attribute of this table is the payment date. Another attribute is a foreign key to the case dimension described below.
2.) DIM_Case: A dimension table showing information on each case, including the case start date. DIM_Case
3.) DIM_Date: A date dimension table.
(For added clarification: The FACT_Payment payment date has to be 5, 10, 15 months etc... after the DIM_Case start date.)
Any ideas, comments, experience with something like this?
I'm working on my first data warehouse and I'm not sure how I should name the columns in the database.
The first phase of the data warehouse is to store a bunch of data from one third party source. The source contains over 100 pieces of data and the business user doesn't even know what some of the fields are but he wants to store everything. The third party refers to the each field with a somewhat cryptic short name and a longer description. The short name isn't always cryptic.
My question is am I better off naming my columns the same as the source system's short name so that I can easily debug problems later? Should I instead try to shorten their definition into something meaningful? On a side note, I'm 100% positive that we'll never populate the tables in questions with data from an additional source.
I need someone to give me a small briefieng on how a enterprise data warehouse (EDW)differs from a regular database? Currently we have an appplication that accesses a database with about 18 tables. We also have a Data Warehouse. For some reason I was thinking that it would be possibly to migrate thedatabase into the data warehouse. The reason is that in looking at the schematic design forthe data warehouse there are some data tables that could also be used by our application that uses the DB. I guess I am confused because I am not sure if a data warehouse is used in the same wayas a database?
i want to create a data mart from an existing OLTP database. for example northwind or i will create an OLTP database. i dont know how i can create data mart from OLTP database. i want to learn that step by step. help me??? please!!
Why shrinkfile empty file does not redistribute data evenly in the primary file group with multiple files:
Please run the script attached to see what the end result is.
This is what I set up last night on my test machine.
1) Create database [FGTest] size 200MB 2) Create table called TEST on primary 3) Insert 40MB of data into test 4) Create another file group called temp in primary size 200MB 5) Shrinkfile('FGTest',emptyfile) so that all data is transfered from FGTest into temp file group. 6) Add another 2 files called DATA2 and DATA3. Both are 200MB. 7) We now have 3 empty files that I want data distributed evenly on. FGTest, DATA2 & DATA3 8) Shrinkfile('temp',emptyfile) to move all the data from temp over the 3 file groups evenly
I would expect at this stage to have the following:
FGTest = 13MB, DATA2 = 13MB, DATA3 = 13MB
(40MB of data over 3 files should be about 13 MBish in each file)
What I actually end up with is this:
FGTest = 20MB DATA1 = 10MB DATA2 = 10MB
It looks as though SQL Server is allocating 50% of all data to the original file and then 50% evenly over the remaining files in PRIMARY.
We have a large 'History' database that is currently about 4.5TB, with most of that in a datafile that is 4.2TB. We wanted to stop growth on the one large data file and have SQL Server allocate new data to the other data files, but this throws an error when we attempt to change the MAXSIZE settings:
ALTER failed for Database 'History' MODIFY FILE failed. Specified size is less than or equal to current size.
The SQL Server is saying we can have a max size of 2TB, and anything over that is blocked. Since this is being blocked, the file continues to grow.
Is there any way to cap the growth of the 4.2TB file and not allow any more data to be written to it?
