September 22, 2015 at 1:02 pm
erichansen1836 (9/22/2015)
Lots and Lots of big words, no links to anything external.
Most everywhere that uses an open source solution that includes other open source projects are published on the internet.
What other open source project links?
This is an undocumented database solution(s) which I have invented myself (Joint Database Technology, Reduction Database Technology). New methodologies, using existing database technologies (MS-Jet Engine/ODBC Databases; Fixed-Length Text File Databases/Perl SDBM Databases).
Microsoft did not even think about using MS-Jet Engine/ODBC this way (10s,100s of MDB files acting as partial tables instead of databases in and of themselves).
You sound like you work for Wikipedia. Do you?
Do you have any proof of this statement or is it just your opinion?
Microsoft did not even think about using MS-Jet Engine/ODBC this way (10s,100s of MDB files acting as partial tables instead of databases in and of themselves).
September 22, 2015 at 1:04 pm
erichansen1836 (9/22/2015)
If for example you wanted to restore the database to yesterday at EOB, first restore your backup off the thumbdrive(s) from yesterday at EOB.
But I don't have a backup from the exact moment I want to restore to, that's the whole point of having the ability to restore to point-in-time.
I backup the whole database once a week (lets say Midnight Sunday). Then I take a differential backup daily (lets say midnight). I also take a Log backup every 10 minutes.
At 10:24:03 some twit deletes half the customers by mistake (notwithstanding, for the sake of this example, that referential integrity would not permit this). A more realistic example would be someone running an update that inadvertently changed more data than was intended.
This is not noticed until 10:31
I restore the Full, Differential and all log backups to a new, temporary, database. When restoring the 10:30 log backup I use the STOPAT parameter set to 10:24:03 (having determined that was when the accident occurred from the data, audit trails, etc.)
I now type
INSERT INTO Production.dbo.Customers
SELECT *
FROM Restored.dbo.Customers AS SRC
WHERE NOT EXISTS
(
SELECT *
FROM Production.dbo.Customers AS DST
WHERE DST.ID = SRC.ID
)
and the problem is fixed.
Or you could restore the database to a certain date/time for today, or you could exclude all changes made prior to 10AM, but keep changes made after 10AM, etc.
Good luck with your referential integrity with that manoeuvre!
Someone has sugegsted that SQL SERVER databases can keep on running and end-users keep on SQL querying/updating the database while a backup/restore is in progress. FINE.
You can't do that with MS-Jet Engine/ODBC-enabled databases. All users must be logged out of the database.
I expect there are some places where noone is working overnight, and a Full backup can be taken at that time without inconvenience etc. That's not true for my databases, they are all used 24/7. Maybe "not much" in the middle of the night, but people can and do use them ... because they can.
BUT ... even if the Full Backup is done in SINGLE USER mode it is absolutely critical that a Log Backup can be taken with the database "running". The log backup is critical for:
Minimal data loss WHEN (not IF ...) the hardware fails, the building burns to the ground, a plane/meteor/whatever hits the building. This has two subordinate options:
1. Restore to the most recent 10-minute log backup. Maximum data loss is 10 minutes.
2. Take a Tail backup. This assumes that the drive that the database is on is toast, but the (separate drive, controller, etc.) which the Log is on is still alive. Thus the log can still be backed up, and combined with a restore (as above) there will be NO data loss.
The ability to restore to point-in-time for an "accident" like the one I described above
The ability to recover from a database corruption. Because the data file and log are maintained separately then if the database becomes corrupted you can restore from a known-good Full Backup (if necessary from days, weeks, months or even years ago). Then restore all the Log File backups since. That leaves you with a clean, corruption free, database.
I cannot comprehend a huge databases where the ONLY backup file that you can recover to is EOB yesterday - after everyone went home so that the database could be taken offline for backup. I just cannot believe that companies that have a database that big would find it acceptable to be told they have to repeat a whole days work in the event of a problem. I do not have a single client, any more, that even has the ability to repeat a day's work. It was never "great", in the old days when that was the only option, from a man-power perspective, but these days there is no paper-trail that exists which could be used to recreate the data, it is all Phone Calls and Emails and so on. It is that way because it CAN be that way. Far more convenient for the users but, quid pro quo, the database must be capable of as close to zero data loss as is possible.
I do not believe that a once-a-day backup of a huge-sized database with many many users is acceptable to any organisation (assuming that they know that that is the case) - unless the database is read only.
September 22, 2015 at 1:04 pm
David Webb-CDS (9/22/2015)
If you haven't actually implemented this with dozens or hundreds of MDBs, why don't you give that a try and then come back and tell everyone how it went? I'd be really interested in how you handled the inevitable restore/recovery situations. And please keep track of the cost so we can actually tell whether this Frankensteinian conglomeration really costs less than a commercial product that would handle the same load. Until then, I wouldn't touch this thing with a 10 foot pole or let you build it in any environment for which I had responsibility.
And for the record, as of right now, I fully agree with David Webb.
Your solution is a half-baked idea that has nothing to safe guard the integrity and validity of the data stored in the *.mdf files.
September 22, 2015 at 1:10 pm
If you haven't actually implemented this with dozens or hundreds of MDBs, why don't you give that a try and then come back and tell everyone how it went? I'd be really interested in how you handled the inevitable restore/recovery situations. And please keep track of the cost so we can actually tell whether this Frankensteinian conglomeration really costs less than a commercial product that would handle the same load. Until then, I wouldn't touch this thing with a 10 foot pole or let you build it in any environment for which I had responsibility.
And then again, I might be wrong ...
David Webb
For one thing, I have done it indirectly with all my experience with ODBC making connections to dozens of Oracle database Servers to gather and consolidate information.
The cost is FREE. It does not cost anything to implement such a database system.
Of course, you are going to have the usual expected costs for any database system implementation such as removable data storage backup devices (like 1 Terabyte thumb drives, or smaller) and the cost of a peer-to-peer network of Windows 7 PCs, etc. if you are starting your business from scratch. You may not need a Windows Server(s). Peer-to-Peer works fine for a small office setting.
If you use the database system such as on a Laptop which you already have, then the cost is ZERO.
I already had my Laptop I used for the Internet.
Totally free to create say a 100 GIG database of 50 *.MDB files (1 Billion Rows).
September 22, 2015 at 1:18 pm
erichansen1836 (9/22/2015)
... The cost is FREE. ...
I guess you don't put much value in the manpower required to implement and maintain this solution.
For best practices on asking questions, please read the following article: Forum Etiquette: How to post data/code on a forum to get the best help[/url]
September 22, 2015 at 1:23 pm
No you haven't. If you've relied on an Oracle back-end to deliver data to you, that has no equivalence to the system you are pitching. Build what you propose. Run it in prod with real users. If they don't beat you senseless in the first couple of months, come back and tell us how it really worked, otherwise this whole discussion is worthless speculation and I have better things to do than think about whether I can use a hammer as a screwdriver.
September 22, 2015 at 1:25 pm
I guess you don't put much value in the manpower required to implement and maintain this solution.
Does not require manpower or womanpower.
All you have to do is use ODBC Admin utility to create an empty MDB file, use ODBC/SQL syntax in a script to create your tables and indexes you want and plan to use, use Windows Explorer to make as many copies of the empty *.MDB file as you need, then load your data to these tables with ODBC/SQL commands in a script. Ready to go. No GUI user interface is needed if you don't want it. You can run batch updates or queries instead if you are that lazy and don't want to put in a couple hours programming effort.
September 22, 2015 at 1:27 pm
Alvin Ramard (9/22/2015)
erichansen1836 (9/22/2015)
... The cost is FREE. ...I guess you don't put much value in the manpower required to implement and maintain this solution.
Yep, Time has a cost. How much is the time worth used building a database system from scratch using free tools versus using that time for other more import business items, like building a client/customer base.
Just because you can do something doesn't mean you should. The is cost involved even if the tools are free. That just happens to be one of the problems that some business don't learn until it is too late.
September 22, 2015 at 1:31 pm
erichansen1836 (9/22/2015)
I guess you don't put much value in the manpower required to implement and maintain this solution.
Does not require manpower or womanpower.
All you have to do is use ODBC Admin utility to create an empty MDB file, use ODBC/SQL syntax in a script to create your tables and indexes you want and plan to use, use Windows Explorer to make as many copies of the empty *.MDB file as you need, then load your data to these tables with ODBC/SQL commands in a script. Ready to go. No GUI user interface is needed if you don't want it. You can run batch updates or queries instead if you are that lazy and don't want to put in a couple hours programming effort.
... All you have to do is ...
That doesn't happen by itself! I'm sure the maintenance doesn't happen by itself either.
For best practices on asking questions, please read the following article: Forum Etiquette: How to post data/code on a forum to get the best help[/url]
September 22, 2015 at 1:37 pm
erichansen1836 (9/22/2015)
I guess you don't put much value in the manpower required to implement and maintain this solution.
Does not require manpower or womanpower.
All you have to do is use ODBC Admin utility to create an empty MDB file, use ODBC/SQL syntax in a script to create your tables and indexes you want and plan to use, use Windows Explorer to make as many copies of the empty *.MDB file as you need, then load your data to these tables with ODBC/SQL commands in a script. Ready to go. No GUI user interface is needed if you don't want it. You can run batch updates or queries instead if you are that lazy and don't want to put in a couple hours programming effort.
If it requires no manpower or womanpower, who creates the system? Who writes the queries? My time has value, a cost. That is why my employer pays me for my time.
You are living in a utopian world. I am not, and most of the people on this site, are not going to spend their time reinventing the wheel when Microsoft is providing us with a fleet of high performance vehicles without us having to write a single line of code to access and support what ever database solutions we may implement and/or support. If we need Point in Time Recovery, we have it. If we need high availability, we have it. If we need replication, we have it. If we need a DR solution, we have it. All of this you still need to write to implement and then spend time supporting it.
You haven't shown us any reason to use what you propose. And constantly saying it is free doesn't make it so. There is no such thing as a free lunch or a free kitten. There are always costs involved.
September 22, 2015 at 1:41 pm
Laptop...Thumbdrive...Golf Cart...32bit...workaround...Free..."connections are only open for a second or two to retrieve a row or write a row"
a few of the keywords i'm picking up here directly telling towards the target audience. I'd be proud of something like that if i built it myself, but i'd also recognize it's limitations.
if you are running a business off a laptop, even if it's a lot of data, i can understand the desire for free, but that's a narrow picture.
with some simple scripting, I could migrate your databases into a single SQL server database in a long afternoon, and KNOW that afterward, it is being backed up, restorable, EVEN TO A SPECIFIC POINT IN TIME if needed.I can compare changes form yesterday to today. I can undo peoples mistakes.
After that, I can spin up a test environment easily. i could then spend the time adding indexes and tuning for queries that actually access the data.
As a Pretty good sized business, When my boss buys a quarter million dollars worth of servers and SANs to store our data on, we are expecting business class performance and reliability, backup restore and redundancy, as well as tried and true technologies.
As David mentioned, There's no way anyone with experience in business applications would even consider the idea.
you skipped over answering ever time anyone mentioned what if something happens to you, what are the backup and restore capabilities.
if I get struck by lightning, a decent consultant or DBA can be hired to pickup where i left off. Sadly, since i follow best practices and all that, someone could swoop in and run with it pretty easily.
To a large degree, a good DBA does his job by making himself obsolete as far as day to day operations go, so he can concentrate on other issues.
Lowell
September 22, 2015 at 2:11 pm
No you haven't. If you've relied on an Oracle back-end to deliver data to you, that has no equivalence to the system you are pitching. Build what you propose. Run it in prod with real users. If they don't beat you senseless in the first couple of months, come back and tell us how it really worked, otherwise this whole discussion is worthless speculation and I have better things to do than think about whether I can use a hammer as a screwdriver.
David Webb, you are here on your own free will unless you are somebody's slave?
You are free to stay or free to go and unsubscribe.
The reason you stay is merely to spread your infectious disease amoung your Trolls who have followed your lead to post nonsense instead of constructive comment.
I order you to stay on this thread and continue making posts.
I ORDER YOU TO STAY.
September 22, 2015 at 2:16 pm
you skipped over answering ever time anyone mentioned what if something happens to you, what are the backup and restore capabilities.
if I get struck by lightning, a decent consultant or DBA can be hired to pickup where i left off. Sadly, since i follow best practices and all that, someone could swoop in and run with it pretty easily.
This is all documented on Microsoft website.
The JetComp.exe utility has been around a long time.
If you have any more questions regarding backup and restore and compact/repair,
I invite you to visit Microsoft's website and some of its partners like
EverythingAccess.com to name 1 of many.
This is simple DBA work a Developer can do, or a semi-technical person in any company department such as accounting or marketing, etc.
Are you unsure of how to stick a thumb drive into a USB port and use Windows Explorer to copy files?
Someone close by should be able to assist you.
If you wish to implement Text File databases using Perl SDBM databases as indexes, you can take what I have told you and google Perl forums.
September 22, 2015 at 2:23 pm
Access databases (.mdb at least, not totally sure about .accdb) were notorious for corrupting if you didn't treat them nicely. I wouldn't trust an .mdb for anything that required security, recoverability etc any further than I could throw it. Putting much of anything in an Access DB and thinking it's going to be secure is a recipe for disaster. Want your data stolen? Great, put it in an Access database. Hope you don't mind HIPPA violations and fines, because you'll get them in spades, because users need FULL access to the folder containing the database.
I guess you could use this kind of technology for larger databases, but why? One violation would no doubt be more expensive than a SQL Server license. (Heck, you could even use a FREE version of SQL Server!!)
The problem with Access databases is that they have no security in the newer versions. It's all in SQL Server...
But if you want to expose yourself to serious security risks with your data, go right ahead.
September 22, 2015 at 2:47 pm
erichansen1836 (9/22/2015)
No you haven't. If you've relied on an Oracle back-end to deliver data to you, that has no equivalence to the system you are pitching. Build what you propose. Run it in prod with real users. If they don't beat you senseless in the first couple of months, come back and tell us how it really worked, otherwise this whole discussion is worthless speculation and I have better things to do than think about whether I can use a hammer as a screwdriver.
David Webb, you are here on your own free will unless you are somebody's slave?
You are free to stay or free to go and unsubscribe.
The reason you stay is merely to spread your infectious disease amoung your Trolls who have followed your lead to post nonsense instead of constructive comment.
I order you to stay on this thread and continue making posts.
I ORDER YOU TO STAY.
Really, you are call us trolls? You, who have steadfastly refused to answer pertinent questions? There is only one true troll on this thread and that is the OP.
You wanted a discussion about your solution. Great, let's have a discussion. That means you also have to answer questions that you are obviously afraid to address, or unwilling to address, or simple have no real idea about. This means listening and accepting, not the same as agreeing to, points of view that are against your solution.
It also means staying on topic, no tangents to the Apollo Project, or golf carts and such. You stay on topic, answer questions, we will to.
The problem is you don't want a true discussion. You want people to see your solution and proclaim your greatness. The problem is, you came to a site where database professionals live and as I keep telling, are job is to protect and defend the data that support the business.
You have avoided all the questions that have been asked regarding just this, defending and protecting the data. Backup and restore processes. Guess what, I can run full, differential, and transaction log backups with users active on the database. I don't have to shut down the database to do it. Depending on the database and how it is setup, it could be possible to allow users on to the database while still restoring other parts of the database. This takes planning to accomplish both in regard to database design and backup procedures. Most of the time, however, no users can be on the system while restoring a database.
High availability. With right version of SQL Server there is database mirroring, always on availability groups, SQL Server clustering. Some of these can be used for Disaster Recovery scenarios as well. Need a separate reporting database? Always on availability, replication are just two possibilities.
What happens if I get hit by a bus? Simple, the company hires another SQL Server DBA. They may not have my institutional knowledge, but they could easily step in and support SQL Server and learn the rest on the job. They don't have to learn some proprietary database access system. Even more difficult if you have to support home grown backup/restore processes, or data replication processes, or what ever else was need to support the business.
Okay, I put a few things out there, your turn. And no copy past from other peoples work. Talk about what you have done, how you handle the things I briefly mentioned.
Viewing 15 posts - 166 through 180 (of 245 total)
You must be logged in to reply to this topic. Login to reply