Achieving High Availability with SQL Server using EMC

Achieving High Availability with SQL Server using EMC SRDF Prem Mehra – SQL Server Development, Microsoft Art Ullman - CSC

Topics Covered Ø Share experiences gained on deploying SQL Server and SAN for a Highly Available Data Warehouse. Emphasis on § § Ø Ø Ø Intersection of SAN and SQL Server Technologies Not on Large Data Base Implementation or on Data Warehouse Best Practices Project Overview Best Practices in a SAN environment Remote Site Fail-over using EMC SRDF and SQL Server Log Shipping

USDA GDW Project Overview Project Client Storage Build Geo-spatial Data Warehouse for two sites with remote fail-over USDA EMC SAN (46 terabytes) Database SQL Server 2000 Implementation USDA / CSC Consultants EMC / CSC / Microsoft / ESRI Geo-spatial Software ESRI Data Management Software

Application Requirements l l A large (46 TB total storage) geo-spatial data warehouse for 2 USDA sites: Salt Lake City & Fort Worth Provide database fail-over and fail -back between remote sites Ø l Run data replication across DS 3 network between sites (45 Mb/sec) Support read- only access at failover sites on ongoing basis

SAN Implementation 1 Ø Ø Understand your throughput, response time and availability requirements and potential bottlenecks and issues Work with your storage vendor § § § Ø Get Best Practices Get design advice on LUN size, sector alignment, etc Understand the available backend monitoring tools Do not try to over optimize, keep LUN, filegroup, file design simple, if possible

SAN Implementation 2 l Balance I/O across all HBAs when possible using balancing software (e. g. , EMC’s Power. Path) Ø Ø Ø l Provides redundant data paths Offers the most flexibility and much easier to design when compared to static mapping Some vendors are now offering implementations which use Microsoft’s MPIO (multi-path IO). Permits more flexibility in heterogeneous storage environments. Managing growth Ø Some configurations offer dynamic growth of existing LUNs for added flexibility (e. g. , Veritas Volume Manger or SAN Vendor Utilities) § Working with SAN vendor engineers is highly recommended

SAN Implementation 3 Benchmarking the I/O System l Before implementing SQL Server, benchmark the SAN. Shake out hardware/driver problems Ø Ø Test a variety of I/O types and sizes. Combinations - read/write & sequential/random Include I/O of at least 8 K, 64 K, 128 K, and 256 K. Ensure test files are significantly larger than SAN Cache – At least 2 to 4 times Test each I/O path individually & in combination to cover all paths Ideally - linear scale up of throughput (MB/s) as paths are added Save the benchmark data for comparison when SQL is being deployed

SAN Implementation 4 Benchmarking the I/O System l l Share results with vendor: Is performance reasonable for the configuration? SQLIO. exe is an internal Microsoft tool Ø On-going discussions to post it as an unsupported tool at http: //www. microsoft. com/sql/techinfo/adm inistration/2000/scalability. asp

SAN Implementation 5 l Among other factors, parallelism also influenced by Ø Ø l For optimizing Create Database and Backup/Restore performance, consider Ø Ø l Number of CPUs on the host Number of LUNs More or as many volumes as the number of CPUs. Could be volumes created by dividing a dynamic disk or separate LUNs Database and Temp. DB Files Ø Ø Internal file structures require synchronization, consider the # of processors on the server Number of data files should be >= the number of processors

Remote Site Fail-over with SQL Server and EMC SRDF USDA Geo-spatial Database

Data Requirements l l l 23 terabytes of EMC SAN storage per site (46 TB total storage) 2 primary SQL Servers and 2 fail-over servers per site 15 TB of image data in SQL Server at Salt Lake City site with fail-over to Fort Worth 3 TB of vector data in SQL server at Fort Worth site with fail-over to Salt Lake City 80 GB of daily updates that need to be processed and pushed to fail-over site

Solution l l l Combination of SRDF and SQL Server Log Shipping Initial Synchronization using SRDF Push updates using SQL Server Log Shipping Use SRDF incremental update to fail -back after a fail-over Use SRDF to move log backups to remote site

Hardware Infrastructure Site Configuration (identical at each site)

Technical Overview – EMC Devices l l l EMC SAN is partitioned into Hyper- Volumes and Meta-Volumes (collections of Hyper-Volumes) through BIN File configuration All drives are either mirrored or Raid 7+1 Hypers and or Metas are masked to hosts and are viewable as LUNs to the OS EMC Devices are identified by Sym Id EMC Devices are defined as R 1, R 2, Local or BCV devices in the Bin File

Technical Overview – Device Mapping Windows Device Manager and SYMPD LIST Output

Technical Overview – SRDF 1 SRDF provides track to track data mirroring between remote EMC SAN devices. BCVs are for local copies. • Track to track replication (independent of host) • R 1 Device is source • R 2 Device is target • R 2 is read/write disabled until the mirror is split

Technical Overview – SRDF 2 Synchronous Mode l Semi-Synchronous with some lag l Adaptive Copy Mode – Asynchronous l Adaptive Copy A – Asynchronous with guaranteed write sequence using buffered track copies Note: only Adaptive Copy A requires additional storage space. All other SRDF replications simply keep a table of tracks that have changed. l

Technical Overview – SRDF 3 • SRDF replicates by Sym Device (Hyper or Meta). • SRDF Devices can be “Grouped” for synchronizing. • SQL Server databases are replicated “by database” or “by groupings of databases” if TSIMSNAP 2 is used. R 1 Group A Database 1 Primary Host Fail-over Host R 1 R 2 R 1 Group B Database 2

Process Overview l l l Initial Synchronization using SRDF in Adaptive Copy Mode (all database files). Use TSIMSNAP(2) to split SRDF group after synchronization is complete. Restore fail-over databases using TSIMSNAP(2) after splitting SRDF mirror. Use SQL Server Log shipping to push all updates to fail-over server (after initial sync). Fail-over database is up and running at all times, giving you confidence that the failover server is working.

Planning l l Install SQL Server and system databases on Primary and Fail-over Servers (on Local non-replicated devices) Create user databases on R 1 devices (MDF, NDF and LDF) on Primary Host Don’t share devices among databases, if you need to keep databases independent for fail-over and fail-back. (Important) Database volumes can be drive letters or mount points

Initial Step Create Databases on R 1 Devices Load Data

Synchronize to Fail-over host 1 l l l Create SRDF Group for Database on R 1 Set Group to Adaptive Copy Mode Establish SRDF Mirror to R 2

Synchronize to Fail-over host 2 l l Wait until Adaptive Copy is “synchronized” Use TSIMSNAP command to split SRDF group after device synchronization is complete. Use TSIMSNAP 2 for multiple databases. TSIMSNAP writes Meta Data about databases to R 1, which is used for recovering databases on R 2 host. Break Mirror Write Meta Data

Attach Database on Fail-over Host l l l Verify SQL Server is installed and running on Fail-over host. Mount R 2 volumes on remote host. Run TSIMSNAP RESTORE command on Fail-over host. Specify either standby (read -only) or norecovery mode. Ø Ø Database is now available for log shipping on fail-over. SRDF Mirror is now broken, but the track changes are still tracked (for incremental mirror and/or fail-back).

Log Shipping – at Primary Site l l Log Shipping volume on separate R 1 device (not the same as the database R 1) Log Backup Maintenance Plan to backup logs to log shipping volume, which is an R 1 device Set R 1 to Adaptive Copy Mode Establish R 1/R 2 Mirror. Logs automatically get copied to R 2.

Log Shipping – at Fail-over Site l l l BCV (mirror) of R 2 Schedule a script that splits and mounts BCV, then restores logs to SQL Server database(s) Flush, un-mount and re-establish BCV mirror after logs have been restored

Process Overview Summary l l Initial Synchronization using SRDF in Adaptive Copy Mode. Use TSIMSNAP(2) to split SRDF group after synchronization is complete. Use SQL Server Log shipping to push updates to fail-over server. Fail-over database is up and running at all times, giving you confidence that the fail-over server is working.

Fail-over Process Fail-over Type Read-only Full Update Required Action No Server Action. Clients would need to point to fail -over server. SQL Command: Restore Database DBName with Recovery

Fail-back Process From Read-only Failover Full Update Failover Required Action None Required. Point Clients to Primary. 1. 2. 3. 4. 5. 6. Run SYMRDF Update command to copy from R 2 to R 1 in Adaptive Copy Mode. Detach database on R 2 after Update Complete. Flush and un-mount volumes on R 2 Run SYMRDF FAILBACK to replicate final changes back to R 1 and write enable R 1 Mount R 1 volumes Attach Database on Primary Host

Closing Observations l l l So far SQL Server 2000 has met High Availability objectives Network traffic across the WAN was minimized, (by shipping only SQL Server Log Copies, once the initial synchronization was completed. ) The dual Nishan fiber-to-IP switches allowed for data transfer at about 16 GB / hour, taking full advantage of the DS 3. This transfer rate easily met USDA’s needs for initial synchronization, daily log shipping, and the fail-back process. The working read-only version of the fail-over database meant that the administrators always knew the status of their fail-over system. The USDA implementation did not require a large number of BCV volumes - as some other replication schemes require.

Closing Observations l l l After the R 1/R 2 mirror has been split, SRDF continues to track updates to R 1 (from normal processing) and R 2 (from log restore process). SRDF is then able to ship only the modified tracks during fail-back or re-synchronization. This process is called an Incremental Establish, or an Incremental Fail-back and is much more efficient than a Full Establish or Full Fail-back. After fail-back, the R 1 and R 2 devices will be “insync”, and ready for log shipping startup with a minimal amount of effort. Since SRDF (initial synchronization, fail-back, and log shipping) all run in adaptive copy mode, the performance on the primary server is not impacted.

Software l l SQL Server 2000 Enterprise Edition Windows 2000 / Windows 2003 Server EMC SYM Command Line Interface EMC Resource Pack

Call To Action l l Understand your HA requirements Work with your SAN Vendor to architect and design for SQL Server deployment Ø Ø l Plan your device & database allocation before requesting a BIN File Decide if sharing devices for databases (use TSIMSNAP or TSIMSNAP 2). Decision effects convenience, space & flexibility of operations Stress test subsystem prior to deploying SQL Server For more information, please email SCDLITE@microsoft. com You can download all presentations at www. microsoft. com/usa/southcentral/

SQL Server Summit Brought To You By: