-
Notifications
You must be signed in to change notification settings - Fork 2
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
[SOLVED] Newly GaleraManager deployed cluster looses/drops/deletes database? (mariadb10.11,Ubuntu22.04.3) #84
Comments
Hi Brian, it appears that GM may drop specifically 'test' database in some circumstances. Could you try a different database name for now? E.g. 'test1'? |
Could you elaborate, what you mean by
Did that happen without your involvement? |
Ayurchen - no I am phrasing it wrong. I myself decided to start testing how smart GM would be - getting cluster up again, so I started stopping service/rebooting servers to see how it acts up. And ended up wondering why that "test" database would just vanish :) Thanks for the heads up that this might be a GM "feature" somehow. I have no problem naming that dummy something else - even if it actually is just a test :D |
Ok, then we need a sequence of your actions to reproduce it - to decide whether it is a bug or we just follow what mysql_secure_installation does. |
Hi
Reproducing should be easy:
VirtualBox: create 4 Ubuntu 22.04 nodes, freshly installed.
Install your GM with 1.11.0 installer on node4.
Connect to GM on node4.
Deploy MariaDB for Ubuntu22.04 (here that ended with MariaDB 10.11.
Do so on all 3 nodes.
No extra settings than 'deploy'
In MySQL on either node:
Create database Test
Create som table/data in 'Test' database.
Then you start experiencing with various restarts of db service/nodes from
cmdline or GM.
I watched the 'Test' disappear multiple times and was sure I did something
very wrong the first few times.
If the database is not called 'Test' it seems to survive.
I did not run any MySQL_secure and did not check if your deploy/recover
does?
Thanks.
…On Tue, 9 Jan 2024, 14.04 Alexey Yurchenko, ***@***.***> wrote:
Ok, then we need a way to reproduce it - to decide whether it is a bug or
we just follow what mysql_secure_installation does.
—
Reply to this email directly, view it on GitHub
<#84 (comment)>,
or unsubscribe
<https://github.com/notifications/unsubscribe-auth/AKVZ2CTZJRRQ7EGIEUJW7VTYNU56VAVCNFSM6AAAAABBS2DTE2VHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMYTQOBTGAZDSNZUGY>
.
You are receiving this because you authored the thread.Message ID:
***@***.***>
|
It seems that this is a "feature" with GaleraManager, that it somehow rips away any databases called "Test". After I renamed my test database to "somethingelse", it is present even after multiple weird attempts to accidentally remove it, lihe what happened to the one called "test" :D Galera - Feel free to explain why "test" needs to go away and when you ensure it does that. I can live with it, but even better when I know the rationale behind it. And thanks anyway for an otherwise great tool!! /Brian |
@brianr-dk can you please test the latest version (gm-installer 1.12.0, Galera Manager GUI 1.8.3). It should have fixed this for you, i.e. the "test" database does not get dropped. |
Thanks a bundle. I'm having issues installing/updating GM, please see the following. First is output from gm-installer install:
And this is from the /tmp/gminstaller.log file:
|
worth noting that Galera Cluster itself changed the signing key, and this is what seems to be failing here. you have an old key. can you start with a fresh install? there are of course ways to update keys, see: https://wiki.debian.org/SecureApt new key announcement re: Galera Manager - https://galeracluster.com/2024/01/galera-manager-january-2024-release/ |
Thanks. I see that somethings changed since my initial install, to the apt-key things in Ubuntu/Debian. I think I've got most of the issues sorted out now, but your install still fails: Attempts to clean up:
Oh, my apt is set up like this - also "as is" from the initial install from a fresh Ubuntu22 download.
So, still no luck really :) |
Solved. It looks as if you've created an installer with the following requirement on at least Ubuntu 22? :
with that added, gm-installer doesn't complain about "ERRO[0011] Failed to complete installation: command failed (stepName=add_apt_repository_universe, commandId=3, commandType=ExecCommand): exec: "add-apt-repository": executable file not found in $PATH" |
Forgot to close :) |
Hi
I finally managed - with GM to create a running cluster on Ubuntu22.04 and was all smiles.
I have 3 nodes and a 4. node to run galera manager on.
Used the GM to deploy Mariadb10.11 on the 3 nodes. No problems. I didn't try educating GM on the settings for anything Galera or MariaDB, just "deploy" and run.
Created a "Test" database and stuffed some data in it. It was replicated to the other nodes. Still smiles.
Then, suddenly I started experiencing the stop/start/restart of the various nodes.
And the "Test" database was gone.
Tried creating it again, and it suddenly deletes itself again.
I am not yet sure exactly WHAT restart makes the database go "puff", but I have tried (I think) every kind of restart sequences I dould think of.
Can anyone suggest what might cause this rather unfortunate feature in Galera?
I haven't seen it before - and really do not need it ;)
Does GM include some parameter during deployment, unbeknownst to me, that makes new databases volatile :)
Best regards
The text was updated successfully, but these errors were encountered: