July 2020 - discuss - lists.mariadb.org

[Maria-discuss] Single-threaded DDL limitation still present in Galera?
by Artem Russakovskii 13 Aug '20

13 Aug '20

Hi, On this page https://mariadb.com/kb/en/mariadb-galera-cluster-known-limitations/, a user in 2017 pointed out the following: 3 years, 3 months ago Björn Schneider <https://mariadb.com/kb/user/id/4916> Schema changes of large tables in Galera <https://mariadb.com/kb/en/mariadb-galera-cluster-known-limitations/#comment…> You should be warned that every DDL statement executed on a Galera cluster will per default BLOCK the complete cluster (not only the table, or even just the database the table resides in)! This is the default "Total Order Isolation" (TOI) mode. The DDL statements can't be killed - once issued, it will run until completed or an error occurs. Issuing e.g. a column-altering DDL statement on a large table will take the complete cluster out of commission until every node has completed the migration. Only some operations (e.g. changing DEFAULT values) are always short-timed and won't interefere with the cluster's opperations. Long-running, cluster-blocking DDL are of course a no-go on a productive system. To resolve this issue, there are several resulutions; Percona, for example, provides a script to "online" migrate a table (pt-online-schema-change), or you can use the "Rolling Schema Upgrade" (RSU) for data-compatible changes. More about that in the Galera documentation. In my opionion, this behaviour should definitely added as "observation" to this page - it's definitely not something you'd expect coming from a "normal" MariaDB/MySQL system. We're considering moving to a cluster environment from a 1 master - 3 slave configuration that has proven inflexible, but such a severe limitation seems unusable in production. Is this still the case in 2020? What is the reasoning behind such an architectural decision, if so? Thank you. Sincerely, Artem -- Founder, Android Police <http://www.androidpolice.com>, APK Mirror <http://www.apkmirror.com/>, Illogical Robot LLC beerpla.net | @ArtemR <http://twitter.com/ArtemR>

2 3

[Maria-discuss] Testing Slave's data integrity
by Richard Reina 25 Jul '20

25 Jul '20

Good afternoon, I have a couple slaves that I have set up over the last year. Is there a way to perhaps compare them to the master on a database or table level to ensure their integrity? Thanks

3 2

[Maria-discuss] MariaDbConnector 1.x -> 2.6, problem updating/inserting Torque objects
by Clemens Wyss - MySign AG 22 Jul '20

22 Jul '20

We are about to update MariaDbConnector from 1.x to 2.6. Having a lot of legacy code we (still) make use of Apache Torque (and Village 1.5.3) "on top" of the connector. Option(s) we had to activate so far was/is: useOldAliasMetadataBehavior=true Problem: Updating/inserting (Torque objects) Records does not work anymore with the new connector. Debugging reveals that it is due to the resultSetConcurrency being CONCUR_READ_ONLY when trying to update/insert a Record. Torque seems to initially select the Record to be updated before effetively updating. Hence the Schema "attached" to the record has CONCUR_READ_ONLY >record.schema().column(i).readOnly() // true ☹ I could patch org.mariadb.jdbc.internal.com.read.resultset. SelectResultSet# getMetaData to return > new MariaDbResultSetMetaData(columnsInformation, options, forceAlias, false ) instead of >new MariaDbResultSetMetaData(columnsInformation, options, forceAlias, true ) but that doesn't really feel "right" 😉 I'd appreciate any hint/advice regarding this issue Thx Clemens

3 4

Re: [Maria-discuss] MariaDbConnector 1.x -> 2.6, problem updating/inserting Torque objects
by Clemens Wyss - MySign AG 21 Jul '20

21 Jul '20

The patch would of course be > new MariaDbResultSetMetaData(columnsInformation, options, forceAlias, ** true ** ) instead of >new MariaDbResultSetMetaData(columnsInformation, options, forceAlias, false ) i.e. forcing "updatable" to true Another note/detail: we make use of the Tomcat db connection pool

1 0

[Maria-discuss] Galera Cluster Behind NAT
by battossai 16 Jul '20

16 Jul '20

Hi, Kindly need help regarding setting up MariaDB Galera Cluster between 2 datacenter. I think i have problem that IST transfer is trying to connect to ip private only, in fact this server behind router. I have DNAT ip public to tcp 4568 to that server, but the donor still want connected to ip private destination. Any help what should i do? Thank you. Regards, Franky

1 0

[Maria-discuss] Building 10.4.13 without PAM
by Chris Ross (cross2) 09 Jul '20

09 Jul '20

Hey there again. So, I just today considered turning off PAM, largely to not have to deal with the setuid auth_pam_tool. We’re not going to be using pam, just internal accounts, so. I added -DPLUGIN_AUTH_PAM=NO and -DPLUGIN_AUTH_PAM_V1=NO to my build options, the same way I had turned off various other plugins I don’t need. But, when running cmake, I now see: CMake Error at cmake/plugin.cmake:283 (MESSAGE): Plugin AUTH_PAM cannot be built Call Stack (most recent call first): CMakeLists.txt:413 (CONFIGURE_PLUGINS) I looked in CMakeOutput.log and CMakeError.log but don’t see anything that looks like related errors. And, I asked it not to build AUTH_PAM plugin, so it complaining that it can’t seems unusual. What did I miss? - Chris

2 6

[Maria-discuss] Specifying install libdir when building from source
by Chris Ross (cross2) 06 Jul '20

06 Jul '20

Good morning. I have a simple question, but haven’t been able to find the answer searching or in cmake -LH output. How do I configure the install paths for libraries? By default, as expected, it installs into /usr/lib, but I want to install into /usr/lib64. Assuming there isn’t an easy “just do that” flag, can I set the library path used to install? I’m much more familiar with autoconf/configure than with cmake. Thanks for your help. - Chris

2 3

[Maria-discuss] why stdout? -> [Note] /usr/libexec/mysqld (mysqld 10.0.20-MariaDB) starting as process 115096
by Reindl Harald 05 Jul '20

05 Jul '20

Aug 6 12:24:24 testserver mysqld: 150806 12:24:24 [Note] /usr/libexec/mysqld (mysqld 10.0.20-MariaDB) starting as process 115096 ... why is this now stdout instead mysql logs? on systemd based distributions that spits now to the global syslog and finally on machines restarting 2 instances every hour for backup reasons it annoys until 10.0.19 or 10.0.20 that was not this way

5 13

[Maria-discuss] Group by pushdown for JDBC or FederatedX
by Dan Hill 04 Jul '20

04 Jul '20

Hi! I'm evaluating if I can use Maria's JDBC connector to join tables from Apache Pinot (they have a prototype jdbc connector). When external tables are used (like JDBC or FederatedX), are group bys pushed down or do these connectors return the raw rows to Maria? I see comments about lack of filter pushdowns for FederatedX. I wouldn't the raw level data coming back from Pinot (there are too many rows to be performant). Thanks! - Dan

2 3