#maria Log v0.1

logs Catalogue

This page loads from the db, refresh as and when. It displays 24 hrs
For older data http://marialog.archivist.info/previous.php, logs go back to 2009-01-18
Some user stats
Table logs
date_time
user
message
2017-09-20 16:39:42
voot
One would expect! I did get some joining, so I think the firewall is ok. Normally the firewall is either yes or no, not maybe or sometimes
2017-09-20 16:40:19
MadMerlin|work
there's keepalive traffic for galera that iirc runs on different ports than the initial sst/ist
2017-09-20 16:40:49
MadMerlin|work
so you might be able to join the cluster and then the keepalive traffic gets dropped and you lose the cluster member
2017-09-20 16:42:53
voot
Interesting. Ok. I'll look into that
2017-09-20 18:06:28
sopparus
i have some ssl issues, how can I log them serverside?
2017-09-20 21:49:31
guardian
hello
2017-09-20 21:49:53
guardian
I'm facing difficulties deleting orphan temporary tables
2017-09-20 21:50:13
guardian
drop table `#mysql50##sql-1526_3a`; doesn't drop the temporary table #sql-1526_3a
2017-09-20 21:50:27
guardian
I described what I did in detail here https://dba.stackexchange.com/questions/183463/unable-to-get-rid-of-orphan-temporary-tables if someone has a clue
2017-09-20 21:53:42
voot
exit
2017-09-20 22:47:53
MariaDBstaff
<wlad> @marko:
2017-09-20 23:11:59
Jeff_
I get an error when I run this: mysqlcheck --auto-repair Error Code: 1064.
2017-09-20 23:12:44
Jeff_
I also get an error when I run this: REPAIR TABLE sale
2017-09-20 23:12:46
Jeff_
The storage engine for the table doesn't support repair
2017-09-20 23:12:47
Jeff_
The storage engine for the table doesn't support repair
2017-09-20 23:16:26
Jeff_
How should I check and repair tables?
2017-09-20 23:20:55
mgriffin
Jeff_: innodb tables don't need repair
2017-09-20 23:22:23
mgriffin
was there a patch for 10.0 master regarding MDEV-12290 ? I don't see it in the 10.0.31/10.0.32 changelog
2017-09-20 23:22:46
mgriffin
i have an older than that 10.0 master with a 10.2 slave seeing similar issues to the bug report i think
2017-09-20 23:27:46
mgriffin
stalled on a "Binlog checkpoint" with bad time stamp
2017-09-20 23:28:42
Jeff_
My table lost all of the rows
2017-09-20 23:28:49
Jeff_
What do you recommend?
2017-09-20 23:29:40
mgriffin
Jeff_: which engine used for the table?
2017-09-20 23:30:36
Jeff_
default-storage-engine = InnoDB
2017-09-20 23:30:48
mgriffin
Jeff_: show create table t1\G
2017-09-20 23:30:53
mgriffin
what is ENGINE=
2017-09-20 23:31:19
mgriffin
ah 10.0.32 is the same re: "#700101" binlog checkpoint :(
2017-09-20 23:31:36
Jeff_
show create table
2017-09-20 23:31:41
Jeff_
0 8 17:30:56 show create table Error Code: 1064. You have an error in your SQL syntax; check the manual that corresponds to your MariaDB server version for the right syntax to use near '' at line 1 0.000 sec
2017-09-20 23:32:16
dragonheart
attention to detail is an important it capability.
2017-09-20 23:32:31
Jeff_
# SAFETY # max-allowed-packet = 16M max-connect-errors = 1000000 skip-name-resolve sql-mode = STRICT_TRANS_TABLES,ERROR_FOR_DIVISION_BY_ZERO,NO_AUTO_CREATE_USER,NO_AUTO_VALUE_ON_ZERO,NO_ENGINE_SUBSTITUTION,NO_ZERO_DATE,NO_ZERO_IN_DATE,ONLY_FULL_GROUP_BY sysdate-is-now = 1 innodb = FORCE
2017-09-20 23:33:02
dragonheart
http://bpaste.net/ for more than one line of text.
2017-09-20 23:33:11
Jeff_
sale CREATE TABLE `sale` ( `SaleID` bigint(19) NOT NULL AUTO_INCREMENT, `CounterCount` bigint(20) DEFAULT NULL, `CustomerID` bigint(20) DEFAULT NULL, `EmployeeID` bigint(19) DEFAULT NULL, `Subtotal` decimal(40,4) DEFAULT NULL, `Tax` decimal(30,4) DEFAULT NULL, `Total` decimal(40,4) DEFAULT NULL, `Comments` varchar(500) DEFAULT NULL, `DateCreated` datetime DEFAULT NULL, `Status` int(11) NOT NULL DEFAULT '1',
2017-09-20 23:33:44
Jeff_
https://bpaste.net/show/d7a2d82dbbc9
2017-09-20 23:35:06
dragonheart
very neat.
2017-09-20 23:36:49
mgriffin
the comment in log_event.h for BINLOG_CHECKPOINT_EVENT= 161 says this is not used by replication, i wonder if master_delay reads timestamp before type
2017-09-20 23:37:41
dragonheart
Jeff_: table-open-cache=256M? that sounds far too big. However if you still want to know about your table loosing all rows, I suggest you use mysqlbinlog to look in your binlog to see what deleted it and when. Or the error log to see something majorly corrupted.
2017-09-20 23:39:32
mgriffin
elenst: maybe this bug was closed early, since the patch only affects timestamp when creating 161, but does not "replicate even from older 10.0/10.1 versions"
2017-09-20 23:39:48
mgriffin
Jeff_: get rid of M on 256M...
2017-09-20 23:40:06
mgriffin
probably error log says something "casting this to a real number"
2017-09-20 23:41:07
Jeff_
Is M way to big?
2017-09-20 23:41:22
MadMerlin|work
do you have 256*1024*1024 tables?
2017-09-20 23:41:38
Jeff_
nope :)
2017-09-20 23:41:50
Jeff_
Did that cause the issue?
2017-09-20 23:42:53
dragonheart
unlikely. show your error log.
2017-09-20 23:43:23
Jeff_
Is this the bin log? mysql-bin.000031
2017-09-20 23:44:20
dragonheart
see in your config, there's a thing that says error-log. that would be the error log.
2017-09-20 23:44:25
Jeff_
Is there a viewer for the bin log?
2017-09-20 23:44:44
dragonheart
for binlogs, mysqlbinlog can decode them.
2017-09-20 23:47:10
Jeff_
Will I loose the data if I reboot the server and hope it fixes it?
2017-09-20 23:47:52
dragonheart
diagnose and solve the error. its likely to happen again.
2017-09-20 23:47:59
mgriffin
you never told us the engine.
2017-09-20 23:48:24
Jeff_
MariaDB
2017-09-20 23:48:31
Jeff_
Is that the engine?
2017-09-20 23:50:00
mgriffin
show create table t1\G
2017-09-20 23:50:05
mgriffin
this will have ENGINE=
2017-09-20 23:50:09
mgriffin
say what it says
2017-09-20 23:50:21
dragonheart
i think rather than innodb=force you mean enforce_storage_engine=innodb
2017-09-20 23:50:31
mgriffin
innodb=force is valid
2017-09-20 23:50:45
dragonheart
ok, new one for me. thanks :-)
2017-09-20 23:50:50
mgriffin
it was used before default-storage-engine=innodb to ensure that innodb starts
2017-09-20 23:51:30
mgriffin
before, unless sql_mode=NO_ENGINE_SUBSTITUTION a wrong sized ib_logfile for example would start with innodb disabled
2017-09-20 23:51:51
mgriffin
and then CREATE TABLE t1 .. ENGINE=InnoDB would result in a myisam table (eg with no FK)
2017-09-20 23:52:01
Jeff_
Here is part of the log:
2017-09-20 23:52:02
Jeff_
https://pastebin.com/swUi6JnF
2017-09-20 23:52:16
mgriffin
if, back then, you set innodb=force, innodb would helpfully halt the server
2017-09-20 23:54:13
mgriffin
Jeff_: which engine is the table?
2017-09-20 23:54:22
Jeff_
how can I tell?
2017-09-20 23:54:58
mgriffin
show create table t1\G
2017-09-20 23:55:00
Jeff_
https://pastebin.com/Mxf1GGY4
2017-09-20 23:55:01
mgriffin
this will have ENGINE=
2017-09-20 23:55:04
mgriffin
say what it says
2017-09-20 23:55:29
mgriffin
Jeff_: great, innodb. alright, does innodb work right now? test with:
2017-09-20 23:55:32
mgriffin
show engine innodb status\G
2017-09-20 23:55:39
mgriffin
if you get a bunch of lines of text, innodb works
2017-09-20 23:57:43
mgriffin
btw error log, just tested, does complain about table cache:
2017-09-20 23:57:45
mgriffin
[Warning] option 'table_open_cache': unsigned value 268435456 adjusted to 1048576
2017-09-20 23:58:04
mgriffin
normally i set it to 8192
2017-09-20 23:58:15
mgriffin
(this is not your problem right now)
2017-09-21 06:37:29
knielsen
mgriffin: I suggest you reopen MDEV-12290, explaining that the bug needs to be fixed in 10.0 since it affect the master. And though delayed slave is only >=10.2, a delayed slave can replicate from a 10.0 master
2017-09-21 06:47:27
gokhan
hi folks, I am trying to create multi node galera cluster on ubuntu 16.04 with openstack ansible but I am getting errors with mariadb galera cluster. Second galera node is stucked with pre-sart and It can not be started. Could anyone help me about my problem. I am new to mariadb galera cluster and not experienced a lot
2017-09-21 07:04:49
dragonheart
gokhan: I don't really have time however, a) ensure the galera-cluster-address is set on all nodes https://mariadb.com/kb/en/library/galera-cluster-address/, b) that one and only one node is started with galera_new_cluster(.sh?) c) read the error log for each node.
2017-09-21 09:12:11
sopparus
can I debug ssl issues server side? I have this problem that sometimes a connect can take up to 15 minutes
2017-09-21 09:12:26
sopparus
client doesnt give any errors, and not my server either
2017-09-21 09:12:54
sopparus
its not a load thing, there is nothing going on the server
2017-09-21 09:14:38
jkavalik
dns problem?
2017-09-21 09:15:44
sopparus
dont think so but I can try to reproduce with ip
2017-09-21 09:15:52
sopparus
on server there is skip resolve
2017-09-21 09:16:46
sopparus
could reproduce with ip as well
2017-09-21 09:18:50
sopparus
also forcing tls version and/or cipher doesnt help
2017-09-21 09:18:53
sopparus
clueless:(
2017-09-21 10:27:52
Cocodude
Hello. I've just set up a two-mode Galera cluster, where one node is presently acting as a slave (so the master node elsewhere replicates to the slave which is part of the Galera cluster).
2017-09-21 10:28:26
Cocodude
When the second Galera node connects, I've found that the master-slave replication never catches up, and that I/O appears to be limited to ~200 writes/sec.
2017-09-21 10:28:45
Cocodude
When the second node is not connected, master-slave replication is just fine, and I get many more writes/sec.
2017-09-21 10:28:59
Cocodude
Any ideas why having a second node connecting slows down things so much?
2017-09-21 10:29:09
Cocodude
(the second node is quicker than the first, master node)
2017-09-21 10:52:04
sopparus
jkavalik, here is debug from client https://pastebin.ca/3875667
2017-09-21 10:52:07
sopparus
or last lines
2017-09-21 10:52:10
sopparus
then it just stops
2017-09-21 14:07:06
svoj
serg: ping
2017-09-21 14:09:40
svoj
serg: in MDEV
2017-09-21 15:33:35
serg
svoj: pong
2017-09-21 15:46:29
svoj
serg: I have rough fix in bb-10.3-svoj1, but it generates warnings. So I wonder what are the other options...
2017-09-21 15:46:56
svoj
at least I couldn't find anything reasonable so far
2017-09-21 15:47:41
serg
see the beginning of sql_plugin.h - that's another possible trick to use
2017-09-21 15:48:11
svoj
svoj is looking
2017-09-21 15:49:39
svoj
serg: uhm, SHOW_always_last?
2017-09-21 15:50:32
serg
yeah. but it doesn't have to be SHOW_always_last, you can do, say,
2017-09-21 15:50:46
svoj
serg: I could do something similar, but it looks quite odd
2017-09-21 15:50:54
serg
#define MYSQL_TYPE_TIME2 MYSQL_TYPE_TIME2,MYSQL_TYPE_BLOB_COMPRESSED=140,MYSQL_TYPE_VARCHAR_COMPRESSED=141
2017-09-21 15:50:56
serg
yes
2017-09-21 15:51:13
serg
a proper fix would be to fix mysqlbinlog not to mix server and client code :(
2017-09-21 15:51:57
serg
but, think as you have two hats. Either you put a hat of the server developer on, and then you have no write access to C/C at all
2017-09-21 15:52:17
serg
or you put a hat of the C/C developer on, and then internal server problems don't bother you
2017-09-21 15:52:23
serg
you cannot have two hats at the same time
2017-09-21 15:52:31
svoj
unless I have two heads
2017-09-21 15:52:51
serg
well, yeah :)
2017-09-21 15:52:55
svoj
I kind of like your trick with MYSQL_TYPE_TIME2
2017-09-21 16:15:30
sopparus
are there any mailing lists for mariadb?
2017-09-21 16:15:36
sopparus
cant find anything
2017-09-21 16:15:56
sopparus
oh, found now