Decoux: Geo: what are the limitations of logical backups?
Decoux: Geo: what conversion?
Kriticos: Dumping the data to a single file
Decoux: Geo: is that a conversion?
Baters: So thats the limitation, why would you use that over something else that doesn’t need to do that?
Decoux: Geo: are you sure it’s a conversion?
Amon: Decoux, if its not, the mysql do***entation is wrong
Keveth: See http://dev.mysql.com/doc/refman/5.6/en/mysqldump.html
Decoux: That page doesn’t mention “conversion”
Decoux: Meh, that’s poorly worded.
Decoux: It does extract the table structure from the information_schema, and writes it into sql statements.
Fuss: So back to my question- why would you use that over something else
Johannes: With the other methods, it seems to be just moving the actual files over to another machine or backup, whatever you’re doing. So what makes that more difficult?
Decoux: Geo: try the other approaches and see for yourself.
Bachinski: File and path permissions, ensuring you have all the files and config settings, overwriting the destination setup
Decoux: Geo: make sure you test the integrity of your data after each test.
Urtado: Bachinski: ah. you probably have to take the db down for a small portion of time while you write those files, don’t you
Bachinski: The mysql server has to be stopped for copying and for pasting the files at source and destination, yes
Bachinski: Mysqldump and xtrabackup don’t require that, the table locking and such is managed for you
Bachinski: Geo: anyway, I’m bored with this. try it yourself, there’s no other way to learn, good luck
Aihara: Bachinski: no one is keeping you here : thank you
Decoux: Geo: the onus is on you to make the efforts and tests.
Affeltranger: Hey guys, im running a galera cluster between DC,s. this morning one DC fiber optic cables were pulled up by someone digging the road up, and i had to restart the galera mysql process in the other dc with –wsrep-new-cluster. my question is will this affect the other DC when it comes online as it was also started with –wsrep-new-cluster initially
Rolin: So basically two machines have been started with –wsrep-new-cluster
Milhorn: Sendro: i’d shutdown the isolated nodes and let them sync again when can join the cluster again.
Mulanax: Danblack: problem is i cannot shut them down as the fiber optic link is cut
Barfoot: They will just come up when the cable has been spliced
Bradney: Ive made a database dump anyway.
Hacke: Danblack: i have 2 nodes in UK and 5 in AU. any ideas why the 2 nodes required mysql restarting with –wsrep-new-cluster
Yanosky: Can dump log if interested on pastbin.
Wekenborg: Line 790 is where i shut down and restarted
Xander: Sendro: i’d firewall the isolated nodes off from the reset of the cluster until you can ***ess it. #maria is probably a better channel as galera is outside the normal bounds of Nelles.
Salvia: Thanks danblack, good idea for firewall
Demiel: I’m trying to start mysql, but all i’m getting is:
Demiel: Sep 14 05:35:18 linux-xhfb mysql1784: Starting service MySQL chmod: cannot access ‘/var/run/mysql/mysqld.pid’: No such file or directory
Demiel: Ah, i see now that there are lots of errors in the database
Demiel: How can i force a recovery?
Senner: 2015-09-14 05:39:35 3145 ERROR WSREP: gcs/src/gcs_group.cpp:group_post_state_exchange:319: Reversing history: 0 – 0, this member has applied 37232744 more events than the primary component.Data loss is possible. Aborting.
Parayuelos: Danblack: any ideas how to discard the updates on the outdated notes