#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
2014-07-31 10:59:12
[diablo]
morning
2014-07-31 10:59:25
[diablo]
guys, anyone running the current mariadb cluster RHEL6 packages?
2014-07-31 10:59:30
[diablo]
I get a dep for socat
2014-07-31 13:41:28
[diablo]
mmm I'm evaluating the MariaDB manager... I've created 3 x nodes
2014-07-31 13:41:33
[diablo]
first is OK
2014-07-31 13:41:49
[diablo]
2nd has been started, and is now trying to sync
2014-07-31 13:41:55
[diablo]
(previous attempted failed)
2014-07-31 13:42:04
[diablo]
3rd done probe, but no option now to start
2014-07-31 13:42:07
[diablo]
:-/
2014-07-31 13:42:26
[diablo]
No Command is currently available on this node
2014-07-31 13:44:29
[diablo]
and either I'm missing something but I don't see any documentation except for the API
2014-07-31 16:17:08
davixx_
select @@slave_parallel_threads; ==> 0
2014-07-31 16:17:34
elenst
svar: i don't see it from here
2014-07-31 16:17:50
elenst
it looks like 1698 is trying to do lock table .. write
2014-07-31 16:17:55
elenst
or an analogue of it
2014-07-31 16:18:06
[diablo]
I'm shocked the logs don't give me info
2014-07-31 16:18:14
elenst
which would be blocked by MDL_SHARED_READ
2014-07-31 16:18:16
[diablo]
as to why I can't start a second node
2014-07-31 16:18:19
svar
elenst, yes that was an update on the song table
2014-07-31 16:19:15
[diablo]
my log shows http://pastie.org/9434532
2014-07-31 16:20:10
svar
elenst, looking at script only see show tables and show databases command
2014-07-31 16:21:58
elenst
svar: is it reproducible?
2014-07-31 16:23:12
svar
elenst, that replication as never show such locking issues before upgrading to 10.0 , that slave have no load exepting that script doing show tables and show databases , it appens 2 times on 2 different queries
2014-07-31 16:23:53
elenst
svar: ask them to enable general log
2014-07-31 16:24:45
[diablo]
what time is monty normally around?
2014-07-31 16:24:47
[diablo]
anyone know
2014-07-31 16:24:51
elenst
svar: if 1698 was just executing a single short query, why does it also have a lock on t_album? (lets forget the global read lock for a moment)
2014-07-31 16:25:05
elenst
svar: apparently, it was executing more than just a write on t_song
2014-07-31 16:25:34
davixx_
locked query : UPDATE `believe`.`t_song` SET `id_producteur`=106555 WHERE `id_album`=2774;
2014-07-31 16:27:52
elenst
davixx_: oh sorry, didn't realize you were talking on the same topic :)
2014-07-31 16:29:11
elenst
[diablo]: if you are going to ask about MariaDB Manager RPMs, I doubt you need montywi. But answering your question -- I don't think anyone can give you an exact time window at the moment
2014-07-31 16:29:36
svar
elenst, sorry i should have say that so how the replication can keep a lock on query that is already past
2014-07-31 16:29:58
elenst
svar: davixx_: is it done inside a transaction? can you find it in the binlog and see what else is in that transaction?
2014-07-31 16:30:09
[diablo]
elenst, well it's the official ones.. It's bizarre, any of the 3 x nodes can be started, but starting either of the second two after that always fails
2014-07-31 16:30:52
elenst
[diablo]: almost everything i know about MariaDB Manager RPMs is that montywi is hardly the person to ask about them
2014-07-31 16:31:05
[diablo]
elenst, oh
2014-07-31 16:31:31
[diablo]
elenst, any idea where I can get some advice from then? I signed up for the enterprise manager evaluation
2014-07-31 16:31:39
svar
elenst, ok will do a binlog extract any ftp to upload ?
2014-07-31 16:32:10
elenst
svar: ftp.askmonty.org:/private
2014-07-31 16:32:27
elenst
svar: kolbe: do you know if there is still an IRC channel for MariaDB Manager?
2014-07-31 16:35:00
kolbe
elenst: i'm not sure. not much use in trying to find one, though, at this point...
2014-07-31 16:35:18
elenst
kolbe: well, see the question above
2014-07-31 16:35:35
elenst
elenst is just trying to dump it on somebody who can possibly say anything sensible on the topic
2014-07-31 16:36:57
elenst
svar: please put show full processlist, metadata locks and show slave status also, while you are there
2014-07-31 16:44:59
ujjain
how od i know if I use innodb or xtradb?
2014-07-31 17:02:36
[diablo]
OK well what a waste of a day... sigh... totally gob smacked that the damn official commercial version fails ... not impressed
2014-07-31 17:03:30
[diablo]
"Get Started Today. Only takes 15 minutes ..." #not
2014-07-31 17:10:30
svar
elenst , ok we found something else here pt-table-sync is sometime use to copy some table to a dev server , that may just explain what was observed
2014-07-31 17:11:15
elenst
svar: okay, thanks fo the update
2014-07-31 17:11:20
svar
elenst, tx , we disable it and will see it just coming back
2014-07-31 17:11:26
elenst
okay
2014-07-31 17:28:10
JimmyNeutron
For MariaDB 5.5, which package should I use for SUSE 11 x64 bit? RPM for Red Hat, Fedora, and CentOS?
2014-07-31 17:28:39
JimmyNeutron
or is SUSE 11 and SUSE 10 even supported?
2014-07-31 17:29:03
kolbe
JimmyNeutron: you can use the generic linux tarball
2014-07-31 17:29:28
JimmyNeutron
kolbe, ok..Thanks! I'll test it out on my vm test machine
2014-07-31 17:29:36
kolbe
great idea
2014-07-31 17:49:42
nirbhay
ujjain, look through the comment field of "show engines".
2014-07-31 17:55:27
jplindst
nirbhay: I little bit modified you changes to server, the cmake changes where fine. We need to add protection on wsrep_get_checkpoint to avoid crash if xtradb/innodb is not compiled using WITH_WSREP, i.e. add check that handler functions called from wsrep_xxx.cc are not NULL
2014-07-31 17:57:17
nirbhay
jplindst, ok..
2014-07-31 17:57:19
jplindst
nirbhay: I will create bug for you about that tomorrow, fix on 10.x
2014-07-31 17:57:34
jplindst
nirbhay: not critical
2014-07-31 17:57:38
nirbhay
or 10.1?
2014-07-31 18:01:09
jplindst
nirbhay: better to do for both, to protect users compiling themselfes
2014-07-31 18:01:37
nirbhay
jplindst, sounds good.
2014-08-01 03:35:47
BossR_
does anyone here have experience with MariaDB 10.0 with a PHP Script called Socialengine?
2014-08-01 03:38:48
horia
hello, any input on this would be much appreciated http://pastebin.com/raw.php?i=qzdSDgnC
2014-08-01 03:39:49
horia
using --wsrep-new-cluster
2014-08-01 03:53:53
mgriffin
BossR_: you might get a better response with a more specific question
2014-08-01 03:56:41
mgriffin
horia: strange. do you have selinux or apparmor enabled?
2014-08-01 03:58:01
horia
mgriffin: this is on FreeBSD
2014-08-01 03:58:59
mgriffin
horia: did you install via ports?
2014-08-01 03:59:12
horia
yes, the ports I make
2014-08-01 03:59:37
mgriffin
hm, is this os supposed to be supported with maria55 + galera? i always use linux and never checked
2014-08-01 04:00:01
horia
it's supported
2014-08-01 04:00:40
BossR_
I am just having some read perfoemance issues and I can't find a way to improve performance... its a long story... so it helps if someone has experience all aorund with my situation
2014-08-01 04:00:43
mgriffin
I just didnt see it on https://downloads.mariadb.org/mariadb-galera/5.5.38/
2014-08-01 04:01:16
mgriffin
BossR_: show the table definition and explain. you can try on #mysql if it is slow here
2014-08-01 04:01:40
mgriffin
(since the question may be agnostic anyway)
2014-08-01 04:02:37
horia
mgriffin: http://www.freshports.org/databases/galera/ and I'm porting mariadb-galera
2014-08-01 04:04:38
horia
mgriffin: this is the port https://redports.org/browser/horia/databases/mariadb55-galera-server
2014-08-01 04:04:54
horia
I've added some patches now like http://bazaar.launchpad.net/~codership/galera/3.x/revision/182
2014-08-01 04:05:28
seekwill
horia: http://galeracluster.com/documentation-webpages/gettingstarted.html#system-requirements
2014-08-01 04:05:39
seekwill
Seems like FreeBSD isn't listed there...
2014-08-01 04:07:30
mgriffin
horia: https://groups.google.com/forum/#!topic/codership-team/voWgTVQHP7I
2014-08-01 04:07:39
horia
seekwill: I've been in contact with galera people about this port, and https://launchpad.net/codership-mysql/5.5/5.5.34-25.9 has FreeBSD bin
2014-08-01 04:07:48
seekwill
oh
2014-08-01 04:07:52
mgriffin
horia: if you dont get an answer here maybe Alex last post can help
2014-08-01 04:08:32
horia
I've talked with Shawn a while back
2014-08-01 04:09:37
mgriffin
apparently a company called perceptyx has built it but i couldn't find the build ;)
2014-08-01 04:10:12
horia
yes, perceptix
2014-08-01 04:11:45
mgriffin
horia: you arent running in a jail or anything?
2014-08-01 04:11:56
horia
mgriffin: yes
2014-08-01 04:12:00
mgriffin
horia: sorry the mariadb developers dont seem to be online at the moment, so seekwill and i are guessing
2014-08-01 04:12:08
mgriffin
horia: maybe unjail and see what happens?
2014-08-01 04:16:11
mgriffin
horia: or check praudit
2014-08-01 04:21:44
mgriffin
wow http://www.mysqlperformanceblog.com/2012/12/19/percona-xtradb-cluster-pxc-what-about-gra_-log-files/
2014-08-01 04:42:35
horia
mgriffin: jail's not the problem, I'll afk a bit and thanks for help
2014-08-01 04:43:10
horia
also thanks seekwill, bbl
2014-08-01 04:43:48
seekwill
haha, I didn't do anything :)
2014-08-01 09:16:33
timour
serg, knielsen: Hi!
2014-08-01 09:17:23
timour
serg, knielsen: the office that hosts the two buildbot needs to rearrange the rack and the internal network.
2014-08-01 09:18:05
timour
serg, knielsen: On Monday afternoon-evening EET the machines may not be accessible. They will try not to power them off.
2014-08-01 09:18:26
timour
montywi, -^
2014-08-01 09:18:44
timour
elenst, hi!
2014-08-01 09:18:48
timour
elenst, -^
2014-08-01 09:19:26
elenst
timour: hi! ack, thanks for notifying
2014-08-01 09:22:57
timour
elenst, maybe it makes sense to shutdown the two boxes on Monday until Tuesday.
2014-08-01 09:23:35
elenst
timour: who can do it, and who can bring them back up afterwards?
2014-08-01 09:23:50
timour
elenst, I have to leave, but if there is any problem booting I can go to the office on Tuesday.
2014-08-01 09:24:09
timour
elenst, If they just press the power button, wouldn't the machines just start?
2014-08-01 09:24:19
timour
They will try not to shut them down.
2014-08-01 09:24:35
timour
If nothing helps, I will go there.
2014-08-01 09:25:00
elenst
timour: okay, thanks. I suppose, but I mean, who can press the button, who should we ask. Anyway, lets see, maybe they'll manage to magically do everything without shutting them down
2014-08-01 09:25:01
timour
elenst, my mobile is still the same, its in the wiki
2014-08-01 09:25:09
elenst
so that you don't have to go there
2014-08-01 09:25:17
elenst
timour: okay, thanks
2014-08-01 09:25:24
timour
elenst, they know in the office, they will push the buttons. after all they use MariaDB :)
2014-08-01 09:25:31
elenst
ah
2014-08-01 09:25:34
elenst
that's better :)
2014-08-01 09:26:21
elenst
okay, so we'll assume the machines are down on Monday, and will expect them back on on Tuesday
2014-08-01 09:26:27
timour
elenst, so tell whoever is responsible nowadays. alternatively, they will just pull the plug if needed. Otherwise the machines can be offline for few hours.
2014-08-01 09:26:33
timour
elenst, yes
2014-08-01 09:26:43
timour
elenst, ok, I have to go, best regards to everyone.
2014-08-01 09:26:50
elenst
timour: deal, thanks
2014-08-01 09:26:53
elenst
later
2014-08-01 09:28:06
Thib
Hi guys, quick question : when using Galera, is there a way to sync only one DB ?
2014-08-01 09:29:09
Thib
For example, I am seeing that the mysql DB is being synched, which makes my root password the same across all of my nodes, which is something that I'd like to avoid...