The Copy Cat Poll

  • pjdiller (8/3/2012)


    Main Production DB is around 3 TB, current growth rate is roughly 1 TB / year.

    Wow... An I thought my 20GB reporting database had lots of records in it...

    Ben

    ^ Thats me!

    ----------------------------------------
    01010111011010000110000101110100 01100001 0110001101101111011011010111000001101100011001010111010001100101 01110100011010010110110101100101 011101110110000101110011011101000110010101110010
    ----------------------------------------

  • the number of copies depends. Most have

    prod

    plus 2-3 in dev, test, ripAndtear.

    There are ETL's into a BI area for various systems and reporting.

    Also there is a series of backups both delta and full on a set schedule , some kept on site some off.

    As well I as the DA have copies of some that I use for various reasons periodically like building cubes to check data quality and identify other data issues and possible further integration.

    Not all gray hairs are Dinosaurs!

  • Well, I'm pretty sure ours is full of extra junk. 🙂 tons of auditing and audits of the auditing... goes on and on. It's a good thing disk is just so darn "cheap" right? sure. 😉

  • 1 - Case Mgmt Production D/B with 3 daily full automated backups on disk, plus hourly transactions backup onto disk. Nightly backups to tape. Month end and year end copies to tape.

    1 - Case Mgmt Development D/B, refreshed from time to time from Prod. Nightly backups to tape. No daily scheduled backups. OOPS!! Gonna fix that one right now. Dang!

    1 - Accounting System D/B copy, no automated backups. Nightly backup to tape.

    1 - Case Mgmt System (alternate, small SQL Server DB) backed up nightly to tape.

    Sigerson

    "No pressure, no diamonds." - Thomas Carlyle

  • Here's the setup for my largest app, entire instance setup and it's a high priority app.

    1 prod

    1 QA

    1 dev

    1 SAN snap

    2 FULL b/u .bak files (not compressed)

    12 Diff b/u .bak files (not compressed)

    2 months of tape

    That's 6 SAN copies (not counting the diff .bak's) and about 60 tape copies. So 6, 18, 66, or 78 copies depending on if you think tape b/u's and/or diff's count. This app has it's own instance at about 466 GBs total for all DBs resulting in about 3 TBs of SAN space.

  • We have a few hundred databases spread across our state in multiple data centers. Our standard backup policy (to Tivoli TSM) is a weekly full backup, nightly differentials and hourly transaction log backups. In a few cases the logs are backed up every 15 minutes instead of hourly.

    Most of our databases have just the main production database and the backups.

    About 10-15 systems have test environments and about that many also have DR environments. Some of the ones with a DR would also be HA, but the typical HA solution at the moment is SQL clustering so there is really just the one copy of the database as a shared cluster resource.

    We take DR very seriously, but even more so ever when our Cedar Rapids data center was caught in a "500 year" flood in June 2008.:w00t:

    So, not counting the backups that are stored on a tape system, I would estimate that for 90%+ have just the production copy, 10% have a test environment, 8% have a DR environment and 5% or less have training environments.

    We do not do any development so we rarely have a development copy with the exception for 1 or 2 systems where they really have two test environments: one test environment that matches the current live environment and one that is used by the vendor for testing major upgrades.

    -----------------
    Larry
    (What color is your database?)

  • Our company creates data sets as a product. In QM we have a production database and part of the company production process is running all the data through a few thousand data quality checks. For our QM database set we have the following:

    1 x production server

    1 x development server

    6 x analyst local servers

    3 x pharmacist local servers

    ---------------------------

    11 not including backups.

  • IowaTechBear (8/3/2012)


    We have a few hundred databases spread across our state in multiple data centers. Our standard backup policy (to Tivoli TSM) is a weekly full backup, nightly differentials and hourly transaction log backups. In a few cases the logs are backed up every 15 minutes instead of hourly.

    Most of our databases have just the main production database and the backups.

    About 10-15 systems have test environments and about that many also have DR environments. Some of the ones with a DR would also be HA, but the typical HA solution at the moment is SQL clustering so there is really just the one copy of the database as a shared cluster resource.

    We take DR very seriously, but even more so ever when our Cedar Rapids data center was caught in a "500 year" flood in June 2008.:w00t:

    So, not counting the backups that are stored on a tape system, I would estimate that for 90%+ have just the production copy, 10% have a test environment, 8% have a DR environment and 5% or less have training environments.

    We do not do any development so we rarely have a development copy with the exception for 1 or 2 systems where they really have two test environments: one test environment that matches the current live environment and one that is used by the vendor for testing major upgrades.

    I know what DR is, but what is HA?

    Kindest Regards, Rod Connect with me on LinkedIn.

  • HA = High Availability. 🙂

    Larry

    -----------------
    Larry
    (What color is your database?)

  • Thanks.

    Kindest Regards, Rod Connect with me on LinkedIn.

  • 4 - development, test, production, dataguard (read-only). This doesn't include the cubes for data warehouse which you could arguably say is another.

  • I agree with the Editorial. 7-8 copies, backup excluded, is probably not a low estimate. With 10+ developers inhouse I dare not think of the actual number. I relatively new at my current work and inventing all DBs have just begun. I find new DBs almost on a daily basis. Also a number of ways to maintain them all.

  • Since we support legacy systems and a full on 50 developpers that are redoing all of our systems by phases, we need two sets of development environments. That brings us to a total of 10 environments which are used for coding and/or testing by different users. All are refreshed with the production data 3 to 4 times per year. So they are all as big as production. One environement takes around 100 gigs.

    All of those are backed up on disk and kept for a day and also copied on tape. Tape as a system of archiving that I'm not gonna go into here. Let's just count them a one copy.

    So 10 environnements x 3 disk and tape backups = 30 copies.

    And I'm not counting all the extra databases for in house management or development tools (HR, SCOM, Intranet, TFS, VMWare, MS Project, etc...).

    We're two DBAs and no free coffee 😉

    ___________________________________
    I love you but you're standing on my foot.

  • Our production database is 1TB uncompressed, 500GB compressed. We currently keep 6 copies for various purposes (reporting, development, test, training, etc.).

  • On average my company does the following for a single database:

    1 Production - Transactional

    1 Production - Reporting

    3 copies spread across 3 non-production environments (often data scrubbed but still realistic in space usage).

    1 DR

    7 (minimum) On-site Full backups from the previous 7 days.

    7 (minimum) Off-site Full backups from the 7 days prior to the on-site backups.

    4 data scrubbed copies pulled to local workstations for development (on average).

    Total Average Copy Count: 24

    This is also not an example of one of our active-active environments across geographically separated data centers with federated instances in the cloud either.

    Derik Hammer

    http://www.sqlhammer.com

Viewing 15 posts - 16 through 30 (of 35 total)

You must be logged in to reply to this topic. Login to reply