I am completely new to SQLite and I intend to use it in a M2M / client-server environment where a database is generated on the server, sent to the client as a file and used on the client for data lookup.
The question is: can I replace the whole database file while the client is using it at the same time?
The question may sound silly but the client is a Linux thin client and to replace the database file a temporary file would be renamed to the final file name. In Linux, a program which has still open the older version of the file will still access the older data since the old file is preserved by the OS until all file handles have been closed. Only new open()s will access the new version of the file.
So, in short:
client randomly accesses the SQLite database
a new version of the database is received from the server and written to a temporary file
the temporary file is renamed to the SQLite database file
I know it is a very specific question, but maybe someone can tell me if this would be a problem for SQLite or if there are similar methods to replace a database while the client is running. I do not want to send a bunch of SQL statements from the server to the client to update the database.
No, you cannot just replace an open SQLite3 DB file. SQLite will keep using the same file descriptor (or handle in Windows-speak), unless you close and re-open your database. More specifically:
Deleting and replacing an open file is either useless (Linux) or impossible (Windows). SQLite will never get to see the contents of the new file at all.
Overwriting an SQLite3 DB file is a recipe for data corruption. From the SQLite3 documentation:
Likewise, if a rogue process opens a
database file or journal and writes
malformed data into the middle of it,
then the database will become corrupt.
Arbitrarily overwriting the contents of the DB file can cause a whole pile of issues:
If you are very lucky it will just cause DB errors, forcing you to reopen the database anyway.
Depending on how you use the data, your application might just crash and burn.
Your application may try to apply an existing journal on the new file. Sounds painful? It is!
If you are really unlucky, the user will just get back invalid results from any queries.
The best way to deal with this would be a proper client-server implementation where the client DB file is updated from data coming from the server. In the long run that would allow for far more flexibility, while also reducing the bandwidth requirements by sending updates, rather than the whole file.
If that is not possible, you should update the client DB file in three discrete steps:
Send a message to the client application to close the DB. This allows the application to commit any changes, remove any journal files and clean-up its internal state.
Replace/Overwrite the file.
Send a message to the client application to re-open the DB. You would have to setup all prepared statements again, though.
If you do not want to close the DB file for some reason, then you should have your application - or even a separate process - update the original DB file using the new file as input. The SQLite3 backup API might be of interest to you in that case.
Related
I want to get the path or even the connection string of a sql.DB instance. The package I'm writing doesn't know what the database path is and the application may have multiple database files. However I need to be able to map a specific database to a specific buffered channel, and ignore any additional calls for a database the package has already seen.
The application uses the github.com/mutecomm/go-sqlcipher driver to instantiate the database, if that makes any difference.
Is it possible to discriminate between instances of sql.DB based on the file path of the source database? If so, how do I do it?
The database can be open()ed using the same encryption key and it works fine. Tried with multiple encrypted databases - all can be opened, but not attached.
This works when encrypted and when not encrypted (bytearray is null):
connection.open(file, "create", false, 1024, bytearray);
This only works when not encrypted:
connection.attach("db" + newnum.toString(), file, new Responder(attachEncryptedSuccess, openEncryptedError), bytearray);
Any help is appreciated.
UPDATE:
Just found a strange pattern here:
It seems that if I create an encrypted database, and then create new databases and attach them, everything works fine.
The created files, after unloading, will only be properly opened using the command that they were initially created with. Therefore, the encrypted database that I created before using open() will only open with open() method. All the encrypted databases that were initially created using attach() will only be able to be opened using attach(). It also doesn't matter which database was open()ed first, aka which one is the main database. It can even be not encrypted.
This is something very strange. Is this a bug? Or am I doing something wrong here?
One gotcha that I ran into awhile ago, and it sounds like it might be impacting you. If you are creating both db's from AIR then this should work fine, however if you have created one with any external tool - generally most tools will default the PRAGMA ENCODING = UTF8. AIR, being Adobe, does things a little different than just straight up telling you that they create theirs UTF16-LE.
According to sqlite rules, differing encoding types cannot be attached one way or the other. One way to verify is to use sqliteman or some other sqlite editor to verify the pragma settings.
For me, I ended up having to start from a seeded db (empty databases -just the header- were over written by AIR) that was to be initialized from a template database. If I allowed AIR to create my starting db, it was set to UTF16 to which I could not attach a UTF8 template.
I have a SQLite database that is used by two processes. I am wondering, with the most recent version of SQLite, while one process (connection) starts a transaction to write to the database will the other process be able to read from the database simultaneously?
I collected information from various sources, mostly from sqlite.org, and put them together:
First, by default, multiple processes can have the same SQLite database open at the same time, and several read accesses can be satisfied in parallel.
In case of writing, a single write to the database locks the database for a short time, nothing, even reading, can access the database file at all.
Beginning with version 3.7.0, a new “Write Ahead Logging” (WAL) option is available, in which reading and writing can proceed concurrently.
By default, WAL is not enabled. To turn WAL on, refer to the SQLite documentation.
SQLite3 explicitly allows multiple connections:
(5) Can multiple applications or multiple instances of the same
application access a single database file at the same time?
Multiple processes can have the same database open at the same time.
Multiple processes can be doing a SELECT at the same time. But only
one process can be making changes to the database at any moment in
time, however.
For sharing connections, use SQLite3 shared cache:
Starting with version 3.3.0, SQLite includes a special "shared-cache"
mode (disabled by default)
In version 3.5.0, shared-cache mode was modified so that the same
cache can be shared across an entire process rather than just within a
single thread.
5.0 Enabling Shared-Cache Mode
Shared-cache mode is enabled on a per-process basis. Using the C
interface, the following API can be used to globally enable or disable
shared-cache mode:
int sqlite3_enable_shared_cache(int);
Each call sqlite3_enable_shared_cache() effects subsequent database
connections created using sqlite3_open(), sqlite3_open16(), or
sqlite3_open_v2(). Database connections that already exist are
unaffected. Each call to sqlite3_enable_shared_cache() overrides all
previous calls within the same process.
I had a similar code architecture as you. I used a single SQLite database which process A read from, while process B wrote to it concurrently based on events. (In python 3.10.2 using the most up to date sqlite3 version). Process B was continually updating the database, while process A was reading from it to check data. My issue was that it was working in debug mode, but not in "release" mode.
In order to solve my particular problem I used Write Ahead Logging, which is referenced in previous answers. After creating my database in Process B (write mode) I added the line:
cur.execute('PRAGMA journal_mode=wal') where cur is the cursor object created from establishing connection.
This set the journal to wal mode which allows for concurrent access for multiple reads (but only one write). In Process A, where I was reading the data, before connecting to the same database I included:
time.sleep(0.5)
Setting a sleep timer before a connection was made to the same database fixed my issue with it not working in "release" mode.
In my case: I did not have to manually set any checkpoints, locks, or transactions. Your use case might be different than mine however, so research is most likely required. Nevertheless, I hope this post helps and saves everyone some time!
I am using sqlite in my application only for read access. The DB gets hit often by my application and I could see that the header(100 bytes) of the database is read every time when i access the database.
Precisely speaking, 16 bytes from the 24th byte of the header is read everytime. My question is , if the database is used only for read purpose, why the header is read everytime as the database connection is not closed?..can we make it read it only once?
Thanks!!
Google search gave me this link, and it says
"Your process may promise that it will only read the database, but there
might be some other process writing to it.
Not being a server, sqlite has no other way to find that out than by
reading the header over and over again. It has to check whether the
schema was changed, or whatever other info is in those bytes."
http://www.mail-archive.com/sqlite-users#sqlite.org/msg69900.html
I'm writing an application with a dBASE database file in Borland Delphi 7.
Note: I think this question is file-security related and you can forget the dBASE thing (consider it as a TXT file) in this question.
The database must be accessed just by the application. Then it must be encrypted. Unfortunately dBASE doesn't support any password mechanism and i had to encrypt the file by myself (and i also HAVE to use dBASE)
What approach do you suggest to secure the database file?
The simple one is:
Encrypting the database file and placing it near beside the application EXE file.
When the application runs, it should decrypt the file (with a hard-coded password) and copy the result to a temporary file that has DeleteOnClose and NoSharingPermission flags.
When Closing, application should encrypt the temp dBASE file and replaces the old encrypted file with the new one.
I think this is a fair secure approach. But it have two big problems:
With an undelete tool the user can restore and access to the deleted temp file.
Worse: When application is running, if the system rebooted suddenly the DeleteOnClose flag fails and the temp file remains on hard disk and user can access it.
Is there any solution for, at least, the second part?
Is there any other solution?
You could also try to create a TrueCrypt file-based containter, mount it, and then put the dBase file inside the mounted encrypted volume. TrueCrypt is free (in both senses) and it's accessible via command line parameters from your application (mount before start, unmount before quit).
Depending on what you're doing with the database, you may be able to get away with just decrypting the records you actually need. For example, you could build indexes based on hash codes (rather than real data); this would reduce seeks into the database to a smaller set of data. Each record in the subset would have to be decrypted, but this could be a lot better than decrypting the entire database.