| |
| |
| |
|
Page: 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
Comments:
<0> hi how would you guys implement access levels?
<0> would you make it numbers, or letters
<0> A user could Insert products, a user could use receipts
<1> m0, what do you mean?
<1> it is specific to your application
<1> do you mean access levels for mysql (who should get insert, alter, drop... permissions?)
<1> or are you asking about business logic in an application
<1> i guess it is business logic (after re-reading the question)
<1> i guess it doesn't matter if you give it an id and a name like: 1 = root, 2 = manager, 3 = registered customer, 4 = non registered customer
<1> m0, though this may not be the best place to ask
<1> ;)
<2> hi channel
<1> hi iv26
<2> a quick question - which maximal value an autoincrementing field may take for int and bigint types?
<1> man numeric types
<1> +man numeric types
<1> damnit
<2> thanks
<0> mick_linux sorry, yea numbers like 1- root etc
<3> my log-bin have 100G how i clean it? and start from 0?
<1> http://dev.mysql.com/doc/refman/5.0/en/numeric-types.html
<1> iv26, ^^^
<0> But they would like it to have read permisisons write permissions, etc
<1> i suggest ACLs
<0> what are ACL's ?
<1> access control lists
<1> i don't suggest something like *nix's bit thing that they do w/ numeric values and sticky bits and other nonsencical crap
<0> Hrmm
<1> http://en.wikipedia.org/wiki/Access_control_list
<0> I would guess a field named access_control is suffient
<3> my log-bin have 100G how i clean it? and start from 0?
<0> And I add APWA
<0> something , representing different controls
<0> GNN-Ricardo
<0> try reset master;
<3> what m0?
<3> reset?
<0> !man binary logs
<4> (SHOW MASTER LOGS Syntax) : http://dev.mysql.com/doc/mysql/en/SHOW_MASTER_LOGS.html
<0> reset master; deletes binary logs
<3> i will try, but this is safe?
<0> hmm
<0> I really don't know to tell you the truth
<3> can i turn the bin-log off?
<0> SHOW BINARY LOGS;
<0> see what it says
<0> If it says ERROR 1381 (HY000): You are not using binary logging
<0> it menas your not using binary logging
<0> !man binary-log
<4> (The Binary Log) : http://dev.mysql.com/doc/mysql/en/Binary_log.html
<1> GNN-Ricardo, thats a bad idea
<1> FLUSH LOGS.
<1> to delete them
<1> check out: http://dev.mysql.com/doc/refman/5.0/en/log-file-maintenance.html
<0> GNN-Ricardo in yourconfiguration you can specify how big your log size could be
<0> mick_linux - he is talking about binary logs
<1> you should rotate logs
<0> I guess he is talking about this log /var/lib/mysql/log-bin
<1> there is bin-log rotate script in the comments
<1> he could always rm and then touch the log :)
<3> m0 | fire-bin.000005 | 158678965 |
<3> | fire-bin.000006 | 27975644 |
<3> i have no space in DISK that's why i wanna delete the BINARY LOG
<5> you can have mysql automatically rotate the logs
<5> and just purge binary logs to 'fire-bin.000005' or something
<3> i do not understand
<5> flush logs doesn't delete them
<5> you generally want binary logging turned on
<3> You can delete all binary log files with the RESET MASTER statement, or a subset of them with PURGE MASTER LOGS. See Section 13.5.5.5, RESET Syntax and Section 13.6.1, SQL Statements for Controlling Master Servers.
<5> yes, but you generally want all binary logs that either your slaves are still reading or that have been generated since your last backup
<3> i do not have slave :-)
<5> but you backup your tables?
<3> it's a lineage server, www.la.gnn.com.br
<3> yes every 6 hours
<3> and sent to a NAS BACKUP SYSTEM in another location
<5> good
<5> you want to keep thebinary logs between each backup
<3> i make backup with DUMP, not binary
<5> yes that's fine
<5> but you want the binary logs anyway
<3> why i want?
<5> imagine you backup at noon and you lose data at 5:45 pm
<5> you ahve nearly 6 hours of changes to your data that you lose if you just restore from the backup
<5> with teh binary log you can restore from the backup adn then play the last 5h45m of changes back to the server
<3> if i lose data the server will stop?
<5> and have something that looks pretty much identical to when you lost the data
<6> Easy: Reset your clocks.
<3> how could i lose data?
<5> magic?
<7> bad sector for example
<5> server crash and data corruption? bad query? bad disk?
<3> i have this server have 1 year
<8> on my bsd i often get "write error, disk 105% full" an the like
<3> and i never got bad disk, bad query.
<5> so what?
<5> maybe you'll have a bad disk tomorrow.
<6> GNN-Ricardo: Someone could accidentally rm -rf /
<5> or maybe someone will find a vulnerability in your website and delete all the rows from some table
<9> $qRemove = "DELECT FROM customer WHERE username = '$username' "; is this a correct query?
<5> first it looks like a variable definition
<5> second, what's DELECT?
<3> i wanna know, if the log-bin it's activated and i got the server crash and data corruption the log-bin will fix the problem imediatly?
<5> no, it won't fix the problem immediately. it will make it easier for you to do a full data restore.
<3> or the log-bin it's just a log of all queries that modify the database?
<5> or, for that matter, possible.
<5> it's just a log of queries that modify the db.
<3> i understand
<3> how i restore a log-bin?
<5> mysqlbinlog
<3> i open the bin file and delete the last bugged queries?
<8> DELECT is hot. especially if you forget the WHERE part.
<5> no, gnn
<5> you use mysqlbinlog to decode the binary log
<5> that will show you the queries
<5> and then you play back the queries you need to play back
<3> great!
<5> read "man mysqlbinlog"
<5> it has lots of handy options.
<3> and how i clean the bin-log every 1 month?
<3> or every backup?
<5> look at flush logs and purge master logs
<3> how i execute a query in the mysql command like, mysql -u root -p=x -q=query...
<3> how i execute a query in the mysql command like, mysql -u root -p=x -q=RESET MASTER;
<5> mysql -e "query"
<5> and then your other options
<3> thanks
<5> e.g. mysql -e "flush logs" -uroot
<10> Blargh.
<5> indeed
<10> So I have a master<->master replication pair
<10> with --log-slave-updates or whatever enabled
<10> I'm snapshotting and backing up machine B, and machine B's binlogs
<10> with this setup I can easily restore to machine A and put it back into the cluster with a catastrophic data loss
<10> but I can't figure out how to restore itself (machine B) without screwing up replication on box A
<10> what am I missing? :(
<10> I kinda don't want to snapshot both and keep both binlogs :P
<5> you know about "SET SQL_LOG_BIN=0"?
<10> probably read about it at some point
<10> gleam; yeah, that does as I expect. not sure how it helps my situation?
<10> I get lost in bringing all the binlogs up to date. If I'm saving binlogs from host B and am restoring back to host B. Don't I need to know where host A was?
<10> I'm probably overthinking this :P
<10> for the restore I need to tell host B to not ignore its own updates temporarly. I apply all of the binlogs that I have, then I need to get the very latest updates from host A since host B's binlogs stopped
<10> (the goal of the exercize is obviously to restore either host from backups instead of stressing the active server
<5> oh right, because you've got log-slave-updates on
<10> )
<10> yeah. without it I couldn't pretend to do that trick
<10> I also need to use it so I can create read-only slaves hanging off of the cluster if necessary
<10> and migrate to new hardware platforms without downtime
<5> with log-slave-updates does it show the server id of the server that received the initial insert in the binary log? you could probably patch mysqlbinlog to only apply changes froms erver-id=X
<10> yeah it does. that's not such a big deal though;
Return to
#mysql
or
Go to some related
logs:
rhel requested were marked to be skipped
fglrx forcemonitor
#python
#gentoo
#python
union syntax error at or near select at character
#debian
samba cups ml-2010
No rule to make target libtorrent
porholio
|
|