RMAN ORA-19913: unable to decrypt backup - encryption

I have been sent a RMAN backup, which is dual encrypted, wallet and password.
The Oracle version is 11.1.0.7.
I am trying to restore the database on a new server.
I performed the following steps:
Create database with ORACLE_SID THEDB using a minimal init.ora file.
Configure the Wallet by copying the ewallet.p12 and cwallet.sso files to a directory, and then adding the ENCRYPTION_WALLET_LOCATION entry into sqlnet.ora.
The next steps are:
sqlplus / as sysdba
startup nomount
select status from v$encryption_wallet;
STATUS
----------------
OPEN
quit
Then RMAN:
rman target / nocatalog
set decryption identified by 'the_backup_pwd';
restore controlfile from 'THEDB_CF_xxxxx_yyy';
ORA-19913: unable to decrypt backup
Now, the backup was taken using AES256 encryption, but the default encryption for this empty database is AES128.
I cannot issue the CONFIGURE ENCRYPTION ALGORITHM 'AES256' command in RMAN as it fails with "database not mounted".
I am certain the wallet and password are correct, but still this error occurs.
What step is missing?

We found out that this only happened if the RMAN Encyrption password had special characters like '$' etc.
The only workaround was to use a heredoc with the password hardcoded .. not great ...
rman / target << EOF
set decryption 'ThePa$$Word'
run {
..
}
EOF
Even when run interactively and the password was manually key it we still got the ORA-19913: unable to decrypt backup. It is obviously a Oracle RMAN bug.

Related

How to hide sensitive data from node.conf?

Can someone please give me an example for corporatePasswordStore that is mentioned here:
https://docs.corda.net/node-administration.html?fbclid=IwAR0gRwe5BtcWO0NymZVyE7_yMfthu2xxnU832vZHdbuv17S-wPXgb7iVZSs#id2
I've been doing a lot of research in the last few days on how to hide the plain passwords from node.conf; it's a new topic for me and this is what I came up with so far:
Create a priv/pub key with gpg2
Create a password store with pass (using the key that I generated earlier).
Store all the plain passwords from node.conf inside that password store.
Replace the plain passwords in node.conf with environment variables (e.g. keyStorePassword = ${KEY_PASS})
Create a script file (e.g. start_node.sh) that will do the following:
a. Set an environment variable to one of the passwords from the password store: export key_store_password=$(pass node.conf/keyStorePassword)
b. Start the node: java -jar corda.jar
c. Restart the gpg agent to clear the cached passwords, otherwise you can get any password from the store without passing the passphrase: gpgconf --reload gpg-agent
Pros:
Using the bash file start_node.sh allows to set many passwords as environment variables at once (e.g. keyStore, trustStore, db passwords, RPC user password)
Since we are running the bash file with bash start_node.sh and not source start_node.sh, the environment variable is not exposed to the parent process (i.e. you cannot read that environment variable value inside the terminal where you ran bash start_node.sh
History commands are not enabled by default inside bash scripts.
Cons:
You no longer can have a service that automatically starts on VM startup, because the start_node.sh script will ask for the passphrase for your gpg key that was used to encrypt the passwords inside the password store (i.e. it's an interactive script).
Am I over-complicating this? Do you have an easier approach? Is it even necessary to hide the plain passwords?
I'm using Corda open source so I can't use the Configuration Obfuscator (which is for Enterprise only): https://docs.corda.r3.com/tools-config-obfuscator.html#configuration-obfuscator (edited)
I wrote a detailed article here: https://blog.b9lab.com/enabling-corda-security-with-nodes-configuration-file-412ce6a4371c, which covers the following topics:
Enable SSL for database connection.
Enable SSL for RPC connection.
Enable SSL for Corda webserver.
Enable SSL for Corda standalone shell.
Hide plain text passwords.
Set permissions for RPC users.

Encrypt SQLITE database and use it with FireDac in a windows VCL application

I am writing a Windows VCL desktop using c++ builder. The app uses Firedac and a SQLite database.
I want to protect this SQLite database:
Only my vcl app can connect to it and use it
Other programs and people cannot use the SQLite file
So far I found that the easiest thing that I can do is Encryption, you make it so that anybody without a correct password gets a message like "This file is not a database" if he tries to open it.
Yes we can do this with Embarcadero VCL, doc. using a TFDSQLiteSecurity component and these lines :
FDSQLiteSecurity1->Password = "";/* we are protecting (encrypting) our uncyphered database*/
FDSQLiteSecurity1->ToPassword = 'newpassword';
FDSQLiteSecurity1->ChangePassword();
Unfortunately the VCL Firedac app cannot connect to an encrypted sqlite database so it must decrypt it first.
In order to make things work my program starts and decrypts the database ( removes password) then it connects to the database and use it. When finished I recrypt the database again (when closing app).
THE PROBLEM:
During this time (My program is up and running and the database in uncrypted) can I stop other programs from opening the SQLITE database? Is there a better approach, I am opened for any suggestion
Once it is encrypted by FireDac just set the appropriate properties on the FDConnection. Those would be password to the password and encrypt to the encryption method.
Firedac can open a SQLite database it has encrypted itself so the problem resolves itself by doing that as a better approach. Other programs only see the encrypted version on disk and can't open it without the password.

SQL Server 2016 TDE: Database encryption scan for 'DB_name' was aborted

When I executed Alter database <DB_Name> set encryption on;, the error message such the subject was appeared in Error.log, and encryption_state has been kept 2 (Encryption in progress).
The DB has only few records because it is just test purpose.
Before execution encryption, DMK, Server certificate, DEK was created with no error.
Besides, I executed Select * from msdb.dbo.suspect_pages; to check if some data is corrupted, but no data was corrupted.
How should I do to complete database encryption?
I appreciate if you give some hint.
Thanks,
[Self solved]
1. Delete all certificate and master key
2. Restart the instans
3. Create Key set newly.
4. Encryption again
...then succeeded to encryption.
[TIPS] When TDE is enabled, compatibility should be '130'. Otherwise, encryption scan will be aborted...

SSH Key Permission Denied

I'm trying to set up cloud hosting with Digital Ocean.
Please skip to the bold part with asterisks (***) for the actual problem. Everything below here, above that part is background info.
I need to generate an RSA key pair, so I navigate to my cd ~/.ssh/ directory, then:
ssh-keygen -t rsa
I already have existing id_rsa and id_rsa.pub files, so when prompted:
Enter file in which to save the key (/demo/.ssh/id_rsa):
I enter the following to create a new pair:
~/.ssh/id_cloudhosting
I'm then asked for a passphrase, which I simply press return for "no password":
Enter passphrase (empty for no passphrase):
I repeat the above for confirmation, and the final output looks as follows (just a demo image):
Now that I have two new files, id_cloudhosting and id_cloudhosting.pub I need to copy the contents of the public file to my Digital Ocean hosting 'Add SSH console'. I do that like so:
cat ~/.ssh/id_cloudhosting.pub
Which returns the contents of the file:
ssh-rsa
bUnChOFcOd3scrambledABCDEFGHIJKLMNOPQRSTUVWXYZnowIknowmy
ABCnextTIMEwontyouSINGwithmeHODOR demo#a
I paste the key into my hosting console and it saves successfully.
The next step is where the permission issues start: ****************
I need to "spin up a new server" - step four from their docs. So I enter the following:
cat ~/.ssh/id_worker.pub | ssh root#[my.hosting.ip.address] "cat >> ~/.ssh/authorized_keys"
Which should copy the public key as root to a newly created file called authorized_keys
This step never gets created because I'm immediately asked for a password to my host. I didn't ever create one! I pressed return (or enter) at that point, so I do the same when prompted, and get permission denied!
root#[host.ip.address]'s password:
Permission denied, please try again.
root#[host.ip.address]'s password:
Permission denied, please try again.
root#[host.ip.address]'s password:
Permission denied (publickey,password).
How can I rectify these permission denied issues?
EDIT: FIX BELOW
It seems as though, by using an unconventional (other than id_rsa) file, I needed to explicitly identify the file by doing the following:
ssh root#droplet.ip.address -i /path/to/private_key_file
...be sure not to use the public_key_file there. I am not connected to the server from my terminal. This is after destroying my previous droplet, creating a fresh one, with fresh key files, as #will-barnwell suggested
Assuming you have followed the linked guide up to and through Step Three, when you create a new server from their Web UI use the "Add SSH Keys" option and select the key you added to your account previously.
When actually spinning up a new server, select the keys that you would
like installed on your server from the "Create a Droplet" screen. You
can select as many keys as you like:
Once you click on the SSH key, the text saying, "Your
root password will be emailed to you" will disappear, and you will not
receive an email confirmation that your server has been created.
The command you were using was to add an ssh key to pre-existing server. Judging from the above quote I bet the password that you are being prompted for is in your email.
Why?
When you create a server on Digital Ocean ( or really most cloud hosting services ) a root password is automatically generated for you, unless you set the server up with an authorization key.
Using key authentication is definitely a good security choice, but make sure to read the instructions carefully, don't just copy/paste commands and expect it all to work out.
EDIT: OP's comments on the question have shed additional light on the matter.
New Advice: Blow your server away and set up the SSH keys as suggested, your server is probably unusable if it is not accepting your old SSH key and is prompting you for a password you don't have.
Be careful messing around with your last auth key, add a new one before removing an old one.

RMAN connecting to wrong target

When I run rman target / #mydb nocatalog cmdfile=%commands% msglog=%logfile%, I see connected to target database: OTHERDB (DBID=3786352837, not open).
Help!
suppose your target database SID = HR_VMC, to connect this target database issue the following commands to connect it using RMAN
export ORACLE_SID = HR_VMC
sqlplus
SQL> connect / as sysdba
SQL> select name from v$database;
it shows the HR_VCM
open another terminal and from the comman line issue the following RMAN command
RMAN target /
Setting environment variable ORACLE_DB to the target database helps the cause.
Set ORACLE_SID=mydb
StackOverflow and ServerFault seem to suffer from lack of Oracle expertise. Weak.

Resources