Reputation
14,128
Top tag
Next privilege 15,000 Rep.
Protect questions
Badges
2 20 43
Newest
 Talkative
Impact
~1.1m people reached

Jan
12
answered Slave stopped : Could not parse relay log event entry
Jan
12
comment Slave stopped : Could not parse relay log event entry
The advice from @NawazSohail is not appropriate for the observed conditions, and should not be followed here. You can't blissfully skip over an event of unknown content without consequences, which may not become evident until later.
Jan
11
comment Slave stopped : Could not parse relay log event entry
If you do CHANGE MASTER TO do not use the Read_Master_Log_Pos value. Use Exec_Master_Log_Pos value (95264000) or you'll break replication for sure.
Jan
10
comment Avoid MySQL self-crashing on long semaphore waits
The only time I have seen this, the disk was worse than "slow" -- it was failing -- and because of that, individual reads were still succeeding, but taking an extremely long amount of time. The server isn't trying to crash itself because it's busy -- that is not the intention of the watchdog, as is clear from the response to Bug #47183. The intention is to guard against an actual hang, so changing the timer is likely not the right thing to do -- it should not fire. I suspect you have a disk issue or a bug causing a false alarm. Test the disks?
Jan
7
comment How do I synchronize rows in tables in remote databases?
There are many ways to compare data, such as generating a cryptographic hash (in SQL) of all the concatenated column values and sending that, along with the pk for each row, across the wire to the far end; mismatched rows will have a mismatched hash, and much less data needs to traverse the wire this way, because then only rows with known or suspected mismatches need be transferred on their entirety... but this is an implementation detail you needn't handle, yet, until you solve some much larger problems you likely have not yet identified. This is no minor undertaking.
Jan
7
comment How do I synchronize rows in tables in remote databases?
If the uuid is the primary key, then the "id" column seems superfluous. Your timestamps are microseconds, not milliseconds, but this is only as useful as the accuracy of the respective system clocks, and contrary to your assertion, actually does nothing to prevent simultaneous -- or, more importantly -- conflicting updates. If the records on both systems have changed, why should the one changed first or last be automatically considered the "correct" one. What about deletions? If a row is missing from A, delete it from B? or copy B to A?
Jan
5
comment Why does Table Definitions Opened differ this much between cluster nodes?
One factor could be tables opened for reading, which would only be opened on the server being queried. Another could be writes that were rolled back, since Galera replicates transactions among servers at commit time, not before.
Jan
4
comment Error on connecting to mysqli
The (110) at the end is a code for the specific failure, and 110 means Connection timed out. This is a network connectivity issue, not a MySQL specific issue. TCP port 3306 is (still) not open end-to-end.
Jan
3
revised mysqldump --single-transaction, yet update queries are waiting for the backup
update, minor clarifications, shift speculative stance (heh)
Jan
3
comment MySQL / MariaDB pid ended (cannot start mysql) in CentOS
Also servers.frm, servers.myd, plugin.frm, plugin.myd are probably among the files the log messages indicate aren't found. Several of them are also likely to have a .myi file of the same base name.
Jan
3
comment MySQL / MariaDB pid ended (cannot start mysql) in CentOS
Do you have files called user.frm, user.myd, innodb_table_stats.frm, innodb_table_stats.ibd, innodb_index_stats.frm, innodb_index_stats.ibd somewhere on your system, at this point? If so, where? The lettercase of the extensions may vary, but these files, and several others, should be in /var/lib/mysql/mysql but if they are elsewhere, mention where you found them. You will need to proceed with extreme caution before moving anything around.
Jan
3
comment MySQL / MariaDB pid ended (cannot start mysql) in CentOS
This looks fairly serious. Unfortunately, if you moved the files with the server running, there's only one way to flawlessly recover, and that has to be done using a delicate manual process prior to the server daemon being stopped. The immediate issue, though, is that files that should be in /var/lib/mysql/mysql can't be located.
Jan
3
comment relationships in access
The relationships should reference the primary key of the "ips" table, and the "IP ID" is what should be stored in the referencing columns.
Jan
1
comment error 2013 (hy000) lost connection to mysql server during query while load of mysqldump
Set the value in the [mysql] section of the config, as well? (Not [mysqladmin] though). SELECT @@MAX_ALLOWED_PACKET; to confirm that the server is using these values?
Dec
27
comment MySQL using cpu after cancelling slow query
possible duplicate of How to stop rollback in progress.
Dec
27
comment MySQL using cpu after cancelling slow query
If this was a query that was modifying data, then you now have a runaway rollback on your hands; the server made changes, but kept track of the former data, and is now in the process of permanently reverting everything to the way it was, undoing the changes. For this reason, your best course is to leave it alone. It should eventually finish. Or, you can force it to stop but that's not the necessarily the best or safest course.
Dec
19
comment Why do innodb tables get corrupted so easily?
I wonder if mysqladmin shutdown would be a better shutdown approach than whatever ridiculous thing Windows might be doing... But that doesn't change the fact that you shouldn't need to be shutting down a production server, and when you do, InnoDB should be able to recover, if recovery were needed in the first place, which it shouldn't be...
Dec
18
comment Change MySQL stored procedure 'Database Collation' name
ALTER DATABASE changes the default values. I suspect your "most correct" solution would be to fix this in the dump file, so they are loaded correctly from the start, but an alternative would be to drop and recreate the procedures now that the database itself is as you intended. Does that work if you try it? No other approach comes to mind at the moment.
Dec
18
comment Modify MySQL information_schema (specially create_time) on AWS RDS?
This can't be done on any MySQL server, and is unrelated to RDS. The "tables" in information_schema are not tables at all -- they're internal server data structures represented as tables, for read-only access. Any "why?" is an excellent question from @Phil.
Dec
15
comment With MySQL 5.6, Is there any way to keep a replication slave X seconds behind_the_master?
@ypercube you wouldn't lose the data. It's persisted to the slave's relay log immediately, as standard async replication (as OP has correctly speculated), but the slave sql_thread waits to execute the events until they are MASTER_DELAY seconds in the past. The idea is to provide a recovery window to avoid data loss due to queries that ran on the master at the hands of someone whose fingers are asking to be cut off. The option is poorly named. SLAVE_SQL_DELAY would have been a more intuitive name.