Matlock: Deason: it seems INET6_ATON is ran on the amster, and the binary is written to the binlog
Whybrew: Deason: if it would just make the slaves execute that function my problem would go away
Fabionar: Deason: http://pastie.org/pastes/10424553/text?key=er9lhwaf5ita2li2t8tvg sory I had to mask some values it’s real customer data
Venezia: Slaves executing functions? pah, it’s 2015, use RBR
Venezia: I find it odd that python can’t p**** binary, considering how many db fields are a binary type
Venezia: Seems that’d be a huge problem for everyone
Cologie: Venezia: me too! i lookeda t the issues page etc, no mention of it
Venezia: Maybe some config in python like treat_binary_as_string
Venezia: No idea, i don’t use it, but
Venezia: People here do use it, and i haven’t heard of a problem, but i don’t think anyone uses it to p**** binlogs
Chapek: I don’t think python has any issue with binary i’m sure of it
Schreckhise: But the script is trying to decode the query as utf-8 when it contains binary
Venezia: Ohh, like converting the bytes in the binary into utf8 multibyte characters?
Venezia: Yeah that would break
Venezia: Is it actually using ipv6 or just v4
Gordner: Tbh mostly v4 – but it’s for future proofing
Batesole: In fact, it’s all v4 right now haha
Deason: BlaDe: http://pastie.org/10424562
Deason: BlaDe: what am i doing wrong?
Venezia: Well, short term, you could store the v4’s as an int field. for futureproofing, you could make two bigints to store the v6, but that’s a hacky way to do it. you could also more easily hex encode it in a char field, but that’s both hacky *and* slower
Venezia: I feel the only good solution is to make python realize it’s not a utf8 string and p**** it properly
Pfund: Deason: hmmm. mine shows that too
Thouvenel: Deason: can I search my master binlog events easily for an insert into a table?
Fortuna: Deason: the test case i did looks like yours too
Boxwell: But the python script errors on replication of the insert at the position of the varbinary
Lookadoo: Venezia: I agree 100%
Venezia: BlaDe: http://dev.mysql.com/doc/refman/5.6/en/mysqlbinlog-row-events.html
Venezia: To get a more readable view of binlog events
Deason: BlaDe: so mysql is doing exactly as you hoped, right?
Deason: BlaDe: possibly you just mis-diagnosed the issue?
Busto: Deason: absolutely – mysql is working great. we’ve been replicating binary for years
Labranche: Deason: I just don’t know if I can fix the python thing sql side easily or not
Reyez: Venezia: yeah, originally I was using mysqlbinlog –no-stop but now I’m trying to use the python tool as it saves me some work, if it works.
Neita: Mwillbanks: I only saw your original question and am about to go afk; but did you consider things like filesystem caches?
Skrabanek: Maybe I’m stating the obvious here, but besides query cache there’s a lot of caches everywhere taht could impact your results
Deason: Jmss: connect to 127.0.0.1 and use the % grant instead of localhost, which is unix socket
Deason: DrkCodeman: please do not pm me
Roundtree: Salle: I see, thanks.
Syrett: Deason: If I use 127.0.0.1 instead of localhost, I get error: SQLSTATEHY000 2003 Can’t connect to MySQL server on ‘127.0.0.1’ 111
Deason: Jmss: this way you don’t have to remember to change two p***words
Linnell: What about skip-networking?
Clyatt: Jmss: 127.0.0.1 != localhost in MySQL world
Clyatt: Jmss: 127.0.0.1 connects via TCP/IP. localhost uses unix socket
Gautreaux: Deason: would that be easy for someone who does not know iptables?