Hi,
Please find your error log file and see what error messages it contains. If you can't decide which ones are relevant, please attach it.
Here is the error log when I start the server:
140424 12:39:02 mysqld_safe Starting mysqld daemon with databases from /var/lib/mysql
140424 12:39:02 [Note] InnoDB: The InnoDB memory heap is disabled
140424 12:39:02 [Note] InnoDB: Mutexes and rw_locks use GCC atomic builtins
140424 12:39:02 [Note] InnoDB: Compressed tables use zlib 1.2.8
140424 12:39:02 [Note] InnoDB: Using Linux native AIO
140424 12:39:02 [Note] InnoDB: Not using CPU crc32 instructions
140424 12:39:02 [Note] InnoDB: Initializing buffer pool, size = 128.0M
140424 12:39:02 [Note] InnoDB: Completed initialization of buffer pool
140424 12:39:02 [Note] InnoDB: Highest supported file format is Barracuda.
140424 12:39:02 [Note] InnoDB: 128 rollback segment(s) are active.
140424 12:39:03 [Note] InnoDB: Waiting for purge to start
140424 12:39:03 [Note] InnoDB: Percona XtraDB (http://www.percona.com) 5.6.15-63.0 started; log sequence number 5079025365
140424 12:39:03 [Note] Plugin 'FEEDBACK' is disabled.
140424 12:39:03 [ERROR] Function 'TokuDB' already exists
140424 12:39:03 [Warning] Couldn't load plugin named 'TokuDB' with soname 'ha_tokudb.so'.
140424 12:39:03 [ERROR] Function 'TokuDB_trx' already exists
140424 12:39:03 [Warning] Couldn't load plugin named 'TokuDB_trx' with soname 'ha_tokudb.so'.
140424 12:39:03 [ERROR] Function 'TokuDB_lock_waits' already exists
140424 12:39:03 [Warning] Couldn't load plugin named 'TokuDB_lock_waits' with soname 'ha_tokudb.so'.
140424 12:39:03 [ERROR] Function 'TokuDB_locks' already exists
140424 12:39:03 [Warning] Couldn't load plugin named 'TokuDB_locks' with soname 'ha_tokudb.so'.
140424 12:39:03 [ERROR] Function 'TokuDB_file_map' already exists
140424 12:39:03 [Warning] Couldn't load plugin named 'TokuDB_file_map' with soname 'ha_tokudb.so'.
140424 12:39:03 [ERROR] Function 'TokuDB_fractal_tree_info' already exists
140424 12:39:03 [Warning] Couldn't load plugin named 'TokuDB_fractal_tree_info' with soname 'ha_tokudb.so'.
140424 12:39:03 [ERROR] Function 'TokuDB_fractal_tree_block_map' already exists
140424 12:39:03 [Warning] Couldn't load plugin named 'TokuDB_fractal_tree_block_map' with soname 'ha_tokudb.so'.
Cannot upgrade TokuDB version 24 database. Previous improper shutdown detected.
140424 12:39:03 [ERROR] Plugin 'TokuDB' init function returned error.
140424 12:39:03 [ERROR] Plugin 'TokuDB' registration as a STORAGE ENGINE failed.
140424 12:39:03 [Note] Server socket created on IP: '127.0.0.1'.
140424 12:39:03 [Note] Event Scheduler: Loaded 0 events
140424 12:39:03 [Note] /usr/sbin/mysqld: ready for connections.
Version: '10.0.10-MariaDB-1~trusty' socket: '/var/run/mysqld/mysqld.sock' port: 3306 mariadb.org binary distribution
And here is the error log when I do install soname 'ha_tokudb';
Cannot upgrade TokuDB version 24 database. Previous improper shutdown detected.
140424 12:42:04 [ERROR] Plugin 'TokuDB' init function returned error.
140424 12:42:04 [ERROR] Plugin 'TokuDB' registration as a STORAGE ENGINE failed.
I see that there are some improperly closed databases? How can I repair them? I think these were caused by the service mysql stop failing when doing the release upgrade (removing MariaDB).
Cannot upgrade TokuDB version 24 database. Previous improper shutdown detected.
This is a serious problem.
Currently TokuDB is unable to load the recovery log from a previous version if it wasn't shutdown properly. Here are the options from the TokuDB expert, as posted on #maria IRC channel:
<prohaska7�>�� bring up the old version (probably with clients locked out) and shut it down. this will write a nice shutdown entry into the tokudb recovery log. then bring up the new version (which should see the old log with the nice shutdown and proceed)
<prohaska7> one could create a new instance of the old version, put the tokudb files in its data directory, bring it up and shut it down. then copy the tokudb log files to the new version and bring it up.
<prohaska7> clean shutdown is required.
Ideally, if you have a backup from before the upgrade, just restore it, rollback to the 5.5 version, perform the clean shutdown, and then upgrade normally.
If you don't have a backup, create one before you downgrade, then if you have issues after you have upgraded again, you can use it and only replace the TokuDB log.
If you can bring up a separate instance as the 2nd suggestion above says, it will be safer.
We will discuss the issue with the TokuDB team to see if we can come up with a proper solution in future.
Thanks for reporting.
When I downgraded to 5.5, ha_tokudb.so was missing. I thought it was supposed to be there!
Which 5.5 version did you use, and which package?
I used 5.5.36, from package mariadb-server-5.5
Oh, never mind. I accidentally used the one from the Ubuntu repos
After using the repo deb http://mariadb.mirror.iweb.com/repo/5.5/ubuntu trusty main and installing mariadb-server-5.5, I still get the "Cannot upgrade TokuDB version 24 database. Previous improper shutdown detected" error. Possibly I could try 5.5.34, which is what I was running?
Actually, could you point me to where I could obtain the ha_tokudb.so file for 5.5.34? I think if I use an older version, it might work.
You can use this repo for example: http://ftp.osuosl.org/pub/mariadb/mariadb-5.5.34/repo/ubuntu/
There should be 5.5.34 version on some other mirrors, but probably not on all of them.
The DEB file mariadb-server-5.5_5.5.34+maria-1~saucy_i386 does not contain ha_tokudb.so...I thought it had TokuDB.
TokuDB is present only in 64-bit packages. Is there a reason you are installing i386? If you had TokuDB before, you must be running it on a 64-bit machine.
Oh, oops! I downloaded the 32-bit version! I will see if I can swap out the file...
Agh! The old ha_tokudb is not compatible! What else can I do? I had trouble installing the old version of MariaDB in a virtual machine. Is there another way?
The old ha_tokudb is not compatible with its own log? Does it throw the same error?
Are you certain you had 5.5.34 before, and that ha_tokudb.so has been actually downgraded now?
So far the two options above is all that TokuDB experts could suggest: either downgrade or install the old version separately, copy TokuDB files there, run it, shut it down, copy the log from there.
For the time being, we probably have a better chance to solve the problem of installing the old version than to find another way to fix the TokuDB problem, since the latter doesn't really depend on us, that's the engine specifics.
What kind of trouble have you had installing the old version of MariaDB?
Here is the log when I use the old .so:
140424 16:07:10 mysqld_safe Starting mysqld daemon with databases from /var/lib/mysql
140424 16:07:10 InnoDB: The InnoDB memory heap is disabled
140424 16:07:10 InnoDB: Mutexes and rw_locks use GCC atomic builtins
140424 16:07:10 InnoDB: Compressed tables use zlib 1.2.8
140424 16:07:10 InnoDB: Using Linux native AIO
140424 16:07:10 InnoDB: Initializing buffer pool, size = 128.0M
140424 16:07:10 InnoDB: Completed initialization of buffer pool
InnoDB: Error: log file ./ib_logfile0 is of different size 0 50331648 bytes
InnoDB: than specified in the .cnf file 0 5242880 bytes!
140424 16:07:10 [ERROR] Plugin 'InnoDB' init function returned error.
140424 16:07:10 [ERROR] Plugin 'InnoDB' registration as a STORAGE ENGINE failed.
140424 16:07:10 [Note] Plugin 'FEEDBACK' is disabled.
140424 16:07:10 [ERROR] Can't open shared library 'ha_tokudb.so' (errno: 8, API version for STORAGE ENGINE plugin TokuDB not supported by this version of the server)
140424 16:07:10 [ERROR] Can't open shared library 'ha_tokudb.so' (errno: 8, API version for INFORMATION SCHEMA plugin TokuDB_trx not supported by this version of the server)
140424 16:07:10 [ERROR] Can't open shared library 'ha_tokudb.so' (errno: 8, API version for INFORMATION SCHEMA plugin TokuDB_lock_waits not supported by this version of the server)
140424 16:07:10 [ERROR] Can't open shared library 'ha_tokudb.so' (errno: 8, API version for INFORMATION SCHEMA plugin TokuDB_locks not supported by this version of the server)
140424 16:07:10 [ERROR] Can't open shared library 'ha_tokudb.so' (errno: 8, API version for INFORMATION SCHEMA plugin TokuDB_file_map not supported by this version of the server)
140424 16:07:10 [ERROR] Can't open shared library 'ha_tokudb.so' (errno: 8, API version for INFORMATION SCHEMA plugin TokuDB_fractal_tree_info not supported by this version of the server)
140424 16:07:10 [ERROR] Can't open shared library 'ha_tokudb.so' (errno: 8, API version for INFORMATION SCHEMA plugin TokuDB_fractal_tree_block_map not supported by this version of the server)
140424 16:07:10 [ERROR] Couldn't load plugins from 'ha_tokudb.so'.
140424 16:07:10 [ERROR] Can't open shared library 'ha_tokudb.so' (errno: 8, API version for STORAGE ENGINE plugin TokuDB not supported by this version of the server)
140424 16:07:10 [Warning] Couldn't load plugin named 'TokuDB' with soname 'ha_tokudb.so'.
140424 16:07:10 [ERROR] Can't open shared library 'ha_tokudb.so' (errno: 8, API version for INFORMATION SCHEMA plugin TokuDB_trx not supported by this version of the server)
140424 16:07:10 [Warning] Couldn't load plugin named 'TokuDB_trx' with soname 'ha_tokudb.so'.
140424 16:07:10 [ERROR] Can't open shared library 'ha_tokudb.so' (errno: 8, API version for INFORMATION SCHEMA plugin TokuDB_lock_waits not supported by this version of the server)
140424 16:07:10 [Warning] Couldn't load plugin named 'TokuDB_lock_waits' with soname 'ha_tokudb.so'.
140424 16:07:10 [ERROR] Can't open shared library 'ha_tokudb.so' (errno: 8, API version for INFORMATION SCHEMA plugin TokuDB_locks not supported by this version of the server)
140424 16:07:10 [Warning] Couldn't load plugin named 'TokuDB_locks' with soname 'ha_tokudb.so'.
140424 16:07:10 [ERROR] Can't open shared library 'ha_tokudb.so' (errno: 8, API version for INFORMATION SCHEMA plugin TokuDB_file_map not supported by this version of the server)
140424 16:07:10 [Warning] Couldn't load plugin named 'TokuDB_file_map' with soname 'ha_tokudb.so'.
140424 16:07:10 [ERROR] Can't open shared library 'ha_tokudb.so' (errno: 8, API version for INFORMATION SCHEMA plugin TokuDB_fractal_tree_info not supported by this version of the server)
140424 16:07:10 [Warning] Couldn't load plugin named 'TokuDB_fractal_tree_info' with soname 'ha_tokudb.so'.
140424 16:07:10 [ERROR] Can't open shared library 'ha_tokudb.so' (errno: 8, API version for INFORMATION SCHEMA plugin TokuDB_fractal_tree_block_map not supported by this version of the server)
140424 16:07:10 [Warning] Couldn't load plugin named 'TokuDB_fractal_tree_block_map' with soname 'ha_tokudb.so'.
140424 16:07:10 [ERROR] Unknown/unsupported storage engine: InnoDB
140424 16:07:10 [ERROR] Aborting
140424 16:07:10 [Note] /usr/sbin/mysqld: Shutdown complete
140424 16:07:10 mysqld_safe mysqld from pid file /var/run/mysqld/mysqld.pid ended
I will actually try the .so from 5.5.36, because there is a chance that I had that version (my brain does not remember things very well...)
When I tried to install the old version through apt-get and that repo, something went wrong with the dependencies and apt-get did not want to install them (possibly because they were not available on the repo?)
I still get a similar error log when I use the .so from 5.5.36, which is the latest version I could have used before the upgrade.
There are two problems here.
The server doesn't start because InnoDB fails to start. It happens because of this:
InnoDB: Error: log file ./ib_logfile0 is of different size 0 50331648 bytes
|
InnoDB: than specified in the .cnf file 0 5242880 bytes!
|
It's caused by different defaults in 5.5 and 10.0.
Try to set in your cnf file (in 'mysqld' section) innodb_log_file_size=50331648, it should solve the problem.
The second problem is
140424 16:07:10 [ERROR] Can't open shared library 'ha_tokudb.so' (errno: 8, API version for STORAGE ENGINE plugin TokuDB not supported by this version of the server)
|
It's normal if you are trying to use TokuDB from 5.5 on a 10.0 server or vice versa. Unfortunately, you need to downgrade the whole server (or to install the 5.5 server with 5.5 TokuDB).
I'm using 5.5 server with 5.5 TokuDB. I'm confused.
Okay, the server actually does not fail to start when the config option is set. Yay.
However, TokuDB still fails to load, with the could not upgrade version 24 database error. I will try to swap out the .so again.
I have 5.5.37 installed, and when using the ha_tokudb.so from 5.5.36, I get the same not supported errors.
I fixed it! 
I installed 5.5.36 from the Ubuntu repos, and then copied the tokudb library from the 5.5.36 deb I downloaded from the "official" source. Now I can upgrade!
Thanks for your patience in helping me resolve this issue.
Great!
My apologies for the unpleasant experience. I hope eventually we will be able to get TokuDB be more friendly in terms of upgrade.
I have created an explicit request MDEV-6173 for improving the upgrade logic, and MDEV-6174 for documenting the current situation. Technically MDEV-6173 is a duplicate of MDEV-6165, but I wanted to make a point that it's a general problem, not a user error as it might seem from the current issue. If it gets fixed in TokuDB, or if we find some other way to minimize the risk, this issue will be closed along with MDEV-6173. Meanwhile, I will keep it open.