Tests won't run on our FreeBSD 8.0 machine

[Lee Bieber]

Something appears to be borked on our FreeBSD 8.0 machine (supplied by Ross McFarland) as drizzle doesn't start when running tests. We are running tests ok on our 8.1 machine. For now I have turned off running tests on this machine until this bug is fixed

Using DTR_BUILD_THREAD = 83630
Using MASTER_MYPORT = 8300
Using MASTER_MYPORT1 = 8301
Using SLAVE_MYPORT = 8302
Using SLAVE_MYPORT1 = 8303
Using SLAVE_MYPORT2 = 8304
Using MC_PORT = 8310
Using PBMS_PORT = 8311
Using RABBITMQ_NODE_PORT = 8312
Killing Possible Leftover Processes
Removing Stale Files
Creating Directories
Saving snapshot of installed databases
================================================================================
DEFAULT STORAGE ENGINE: innodb
TEST RESULT TIME (ms)
--------------------------------------------------------------------------------

user_locks.dangling_barrier [ fail ]
drizzletest: Could not open connection 'default': 5 drizzle_state_connect:socket:43

Stopping All Servers
drizzle-test-run: WARNING: Forcing kill of process 83767
Restoring snapshot of databases
Resuming Tests

user_locks.dangling_lock [ fail ]
drizzletest: Could not open connection 'default': 5 drizzle_state_connect:socket:43

Stopping All Servers
drizzle-test-run: WARNING: Forcing kill of process 83810
Restoring snapshot of databases
Resuming Tests

From the error log:

CURRENT_TEST: user_locks.dangling_lock
InnoDB: The InnoDB memory heap is disabled
InnoDB: Mutexes use GCC atomic builtins, rw_locks do not
InnoDB: Compressed tables use zlib 1.2.3
110202 18:01:52 InnoDB: Initializing buffer pool, size = 128.0M
110202 18:01:52 InnoDB: Completed initialization of buffer pool
InnoDB: The first specified data file /home/hudson/hudson/workspace/drizzle-build-freebsd-8.0/
tests/var/master-data/ibdata1 did not exist:
InnoDB: a new database to be created!
110202 18:01:52 InnoDB: Setting file /home/hudson/hudson/workspace/drizzle-build-freebsd-8.0/
tests/var/master-data/ibdata1 size to 20 MB
InnoDB: Database physically writes the file full: wait...
110202 18:01:52 InnoDB: Log file /home/hudson/hudson/workspace/drizzle-build-freebsd-8.0/test
s/var/master-data/ib_logfile0 did not exist: new to be created
InnoDB: Setting log file /home/hudson/hudson/workspace/drizzle-build-freebsd-8.0/tests/var/mas
ter-data/ib_logfile0 size to 20 MB
InnoDB: Database physically writes the file full: wait...
110202 18:01:52 InnoDB: Log file /home/hudson/hudson/workspace/drizzle-build-freebsd-8.0/test
s/var/master-data/ib_logfile1 did not exist: new to be created
InnoDB: Setting log file /home/hudson/hudson/workspace/drizzle-build-freebsd-8.0/tests/var/mas
ter-data/ib_logfile1 size to 20 MB
InnoDB: Database physically writes the file full: wait...
InnoDB: Doublewrite buffer not found: creating new
InnoDB: Doublewrite buffer created
InnoDB: 127 rollback segment(s) active.
InnoDB: Creating foreign key constraint system tables
InnoDB: Foreign key constraint system tables created
110202 18:01:53 InnoDB 1.1.4 started; log sequence number 0
Listening on 0.0.0.0:8350
Listening on 0.0.0.0:8300
Listening on /home/hudson/hudson/workspace/drizzle-build-freebsd-8.0/tests/var/master.sock
/home/hudson/hudson/workspace/drizzle-build-freebsd-8.0/drizzled/.libs/drizzled: ready for con
nections.
Version: '2011.02.2139' Source distribution (drizzle-build-freebsd-8.0)
Note: Forcing kill of process 83810

[Lee Bieber (kalebral-deactivatedaccount)]

Hi Vijay,
Thanks for investigating. I'm including Ross in case he might see something with regards to the port issue on his machine. It does seem strange that it picks 9307 as the master port instead of 9306. However as you noted, when you do a start-and-exit you then are able to connect to drizzle through either port 9306 or 9307

./dtr variablesLogging: ./dtr variables
Drizzle Version 2011.02.2150
Using DTR_BUILD_THREAD = -69.4
Using MASTER_MYPORT = 9307
Using MASTER_MYPORT1 = 9308
Using SLAVE_MYPORT = 9309
Using SLAVE_MYPORT1 = 9310
Using SLAVE_MYPORT2 = 9311
Using MC_PORT = 9317
Using PBMS_PORT = 9318
Using RABBITMQ_NODE_PORT = 9319
Killing Possible Leftover Processes
Removing Stale Files
Creating Directories
================================================================================
DEFAULT STORAGE ENGINE: innodb
TEST RESULT TIME (ms)
--------------------------------------------------------------------------------

main.variables [ fail ]
drizzletest: Could not open connection 'default': 5 drizzle_state_connect:socket:43

Aborting: main.variables failed in default mode.

On Fri, Feb 4, 2011 at 11:41 PM, Vijay Samuel <email address hidden> wrote:

    Hi,
      I did some snooping into the FreeBSD box and I managed to gather some info regarding what is going on, but I m not completely sure as to if that is the reason.

    After doing a ./test-run --start-and-exit and then a netstat -an | grep 9306 and I got this output.

    netstat: kvm not available: /dev/mem: No such file or directory
    tcp4 0 0 192.168.1.45.9306 *.* LISTEN

    conventionally when we do it on orisndriz or my localbox
    vjsamuel@codersville:~/repos/drizzle/test/tests$ netstat -an | grep 9306
    tcp 0 0 0.0.0.0:9306 0.0.0.0:* LISTEN
    tcp6 0 0 :::9306 :::* LISTEN

    Does it have anything to anything with the ip being different and clearly ipv6 is disabled. Im assuming it has to do with the box rather than our code. But on the other hand, it is kindda strange to notice that I was able to pop up a client after doing a start-and-exit and was not able to run the test suite.

    Is MASTER_MYPORT1 an alternative mysql protocol port? if it is then it is a pile of bull shit because i am unable to connect to a client using that port.
    Hope my info was useful.

    Cheers,
     -Vijay

[Lee Bieber (kalebral-deactivatedaccount)]

From Ross:

> What puzzles me most is the ability to do a --start-and-exit and not being
> able to run the test suite. The port is able to listen when i pop up a
> client but is not able to listen when we start up the test suite. Any idea
> on what may be going on?

no clue, but i'm guessing it's something wrong with the testclient,
though i don't know what it would be. i just connected to the test
server via 127.0.0.1 and 192.168.1.45 with drizzle (below.)

i also added a debugging print to the testclient to dump out all of
the connection info:
   name: default, host: localhost, user: root, pass: (null), db:
test, port: 8610

it looks like drizzle and drizzletest use different mechinims to set
up the connection, but may have nothing to do with the problem(s.)

--
-rm

drizzle-build$ ./client/drizzle -h 127.0.0.1 -p 8610
Welcome to the Drizzle client.. Commands end with ; or \g.
Your Drizzle connection id is 1
Connection protocol: mysql
Server version: 2011.02.2137 Source distribution (drizzle)

Type 'help;' or '\h' for help. Type '\c' to clear the buffer.

drizzle> Bye
drizzle-build$ ./client/drizzle -h 192.168.1.45 -p 8610
Welcome to the Drizzle client.. Commands end with ; or \g.
Your Drizzle connection id is 2
Connection protocol: mysql
Server version: 2011.02.2137 Source distribution (drizzle)

Type 'help;' or '\h' for help. Type '\c' to clear the buffer.

drizzle> drizzle-test-run: WARNING: Forcing kill of process 43078
Restoring snapshot of databases
Resuming Tests

filtered_replicator.var_filteredtables [ fail ]
name: default, host: localhost, user: root, pass: (null), db: test, port: 8610
drizzletest: Could not open connection 'default': 5
drizzle_state_connect:socket:43

Stopping All Servers