by

Where communities thrive


  • Join over 1.5M+ people
  • Join over 100K+ communities
  • Free without limits
  • Create your own community
People
Activity
  • 16:45

    philipmarzullo64 on 3.12

    0004415: Ingres Dialect Impleme… (compare)

  • Jun 04 14:28

    erilong on 3.12

    unit tests need H2 (compare)

  • Jun 02 15:15

    erilong on 3.12

    0004412: Add license violation … (compare)

  • Jun 02 15:12

    erilong on 3.11

    0004412: Add license violation … (compare)

  • Jun 02 14:54

    erilong on 3.11

    0004408: Jetty error: KeyStores… (compare)

  • Jun 02 12:59

    erilong on 3.12

    0004411: NoClassDefFoundError d… Merge branch '3.12' of https://… (compare)

  • May 27 13:54

    vanmetjk on 3.12

    0004407: Monitor Email Notifica… Merge branch '3.12' of https://… (compare)

  • May 26 17:42

    chenson42 on 3.12.0

    (compare)

  • May 26 13:22

    chenson42 on 3.11.9

    (compare)

  • May 26 13:09

    chenson42 on 3.10.12

    (compare)

  • May 26 12:50

    chenson42 on 3.9.25

    (compare)

  • May 26 12:40

    chenson42 on 3.8.44

    (compare)

  • May 22 17:05

    erilong on 3.12

    0004225: fix npe converting to … 0004398: Initial load leaves ba… 0004397: Bit Support for Postgr… and 3 more (compare)

  • May 22 14:46

    philipmarzullo64 on 3.11

    0004405: Initial Load: Retrieve… (compare)

  • May 21 16:08

    vanmetjk on 3.11

    0004397: Bit Support for Postgr… Merge branch '3.11' of https://… (compare)

  • May 21 15:27

    erilong on 3.11

    0004398: Initial load leaves ba… (compare)

  • May 21 15:23

    erilong on 3.10

    0004398: Initial load leaves ba… (compare)

  • May 21 15:14

    erilong on 3.11

    0004225: fix npe converting to … (compare)

  • May 21 15:05

    erilong on 3.12

    0004229: [Docs] Invalid column … (compare)

  • May 21 14:56

    erilong on 3.12

    0004226: trigger template varia… (compare)

Marcelo
@marcelosf
i'm still having the same error after making the configuration changes. Now i'm using two nodes and the time out query error persists.
Yu-Phing Ong
@ong.yuphing_gitlab

Why are you use a sql server reserved word as a column name?

it's a legacy database, there are 400 cols in this table (!), I have no way to change the name of that table!

(and don't ask me why there are 400 cols... nobody else can tell me either)
OK I guess I'll just have to try with auto.sync.triggers=false
thanks
joshahicks
@joshahicks
@marcelosf is your installation of SymmetricDS on the same local network as your database? If your database is in the cloud or a remote data center for example you will want to install SymmetricDS close to the database. It does not need to be the same server as the database.
Marcelo
@marcelosf
@joshahicks My SymmetricDS installation and my database is on the same network. SymmetricDS and the databases are installed in diferent hosts.
host 1 => sqlserver 1
host 2 => sqlserver 2
host 3 => SymmetricDS
xShtaket
@xShtaket1_twitter
Hi.

Some introduction – I have 3 SymmetricDS 3.11.7 container that is connected to each own database. Databases have the same structure and works on MySQL. One SymmetricDS is the Master and two another – child. For some list of tables, I created “Load Only Triggers” configuration in sym_trigger (where sync_on_update, sync_on_insert and sync_on_delete have 0 value). Also, I configured router for that table and linked it with triggers in table sym_trigger_router.
I do that configuration because I need sync this tables only via ‘Table reload request’. And to do that I run reload request for all tables that I need in same time.
After configuration I start some little test to make sure that everything works well. In the beginning everything was fine, but then in some cases a faced with some strange issue.
When it works normal, SymmetricDS collected data from all tables that need to be sync and send to my child node via one packet. But sometimes I have a problem, SymmetricDS start send and sync data to child node only one by one table whit huge delay.
I try to understand why it happened, also I try to analyze on which table or what data could be the issue reason, but I can’t, because issue happened randomly.
What I understand and analyzed that in log-file I have this error:

ERROR [master-cloud] [InitialLoadService] [master-cloud-job-15] Error while queuing initial loads StackTraceKey [SqlException:1943736719]
INFO [master-cloud] [DataExtractorService] [master-cloud-extract-default-7] Done with request 7117 to extract table table_name_1 into batches 12398 through 12398 for node 2
INFO [master-cloud] [DataExtractorService] [master-cloud-extract-default-7] Releasing 8 child extract requests that missed processing by parent node
WARN [master-cloud] [DataExtractorService] [master-cloud-extract-default-7] Failed on request 7117 to extract table table_name_1 into batches 12398 through 12398 for node 2
ERROR [master-cloud] [NodeCommunicationService] [master-cloud-extract-default-7] Failed to execute EXTRACT for node 2 and channel reload_queue StackTraceKey [SqlException:3396787583]

And when I restart my Master Node SymmetricDS I have an error like: ‘Lock wait timeout exceeded; try restarting transaction’ and error with table sym_extract_request.
And to fixed that I need to run that SQL update query on my Master database:
UPDATE sym_outgoing_batch SET status = 'IG' WHERE STATUS != 'OK' AND channel_id = 'reload';
UPDATE sym_extract_request SET status = 'OK' WHERE STATUS = 'NE';
After that when I restart my reload request and It’s work normally. But then it happens but again due to a random number of reload request.
And I want to ask if someone faced with the same issue or someone maybe could help me understand how to fix that.

Yu-Phing Ong
@ong.yuphing_gitlab
Does symmetricds support this scenario where a store would replicate a set of tables to/from one source, and another set of tables to/from another source? e.g. store sends one set of sales data to one warehoues, and another (overlapping) set of sales data to another warehouse? I tried to create 2 engine files, but received an error about 2 properties files referencing the same db.user@db.url
joshahicks
@joshahicks
@ong.yuphing_gitlab yes it would still be one source node however you would need two routers and two target groups with two target nodes. The first set of tables would use the router to send the data to target A and the second set of tables would use a different router to send the tables to target B
smolens
@smolens
In sym_data..row_data of the source has control characters HOP and DCS. But when I look at sym_data..row_data the control characters are replaced with "?". The .done file on the source and target has binary,BASE64 and has the control characters as "?". Why are HOP and DCS being replaced "?"?
MSSQL on Linux - Symetric 3.10.4 - Field replicated is VARCHAR(200)
joshahicks
@joshahicks
@smolens whats the charset of the database, table, column
@smolens you could set this parameter to capture all charsets mssql.use.ntypes.for.sync=true
smolens
@smolens
charset for all are set to: SQL_Latin1_General_CP1_CI_AS. I will update that setting and see.
smolens
@smolens
@joshahicks Thanks that worked. sym_data was quite large so it took a while to convert row_data to ntext.
Albrod
@Albrod
@joshahicks Hi, i have a big doubt. Does SymmetricDS know what to do when data_id and batch_id get to the max value ?
In the company i work for, a coworker who whas in charge of symmetricds and db Maintenance in his previous company told me he have wrote a script to renumber those id and reset the sequentials every 6 monts because they run out of ID and SymmetricDS stop working. They work with postgresql 9.6 and Symmetric 2.5
Marcelo
@marcelosf
@smolens I intalled mssql on centos, and now the initial load is working fine. Before that, It was an ubuntu 18.04, I can't figure out why it was not working on ubuntu
smolens
@smolens
@marcelosf I am using Ubuntu 18.04. I started using symmetricds with 3.9 there was an issue with starting symmetricds as a service. I may have had other issues but if I recall correctly the log output helped pin point what I had misconfigured.
Bigbear
@FunnyGoofy
Hi guys, I hope everyone are doing very well
I have a case as shown below
image.png
"id" is primary key in each table. if I want "id" auto-incremental and all records from different "stores" kept other than overwritten in table "corp". is just to configure "Symmetric" possible?
joshahicks
@joshahicks
@FunnyGoofy what we normally do in these situations is create a composite key at corp that includes id and node_id or store_id for example. Then you would have two rows your example. Row 1 : ID=1, STORE_ID=1 Row 2 : ID=1, STORE_ID=2
This would represent the PK so both would be valid and unique. This can be done with a transform that adds a variable transfrom to use the external id from the store node to be added to the STORE_ID col when syncing to corp.
@FunnyGoofy you can also use a column match router if this data needs to be sent bi-directionally from corp to store. This would then use the STORE_ID colunn to determine which store to send the change to. This would result in no changes to the store databases but just the corp database
Bigbear
@FunnyGoofy
@joshahicks awesome, the second one is exact what I want. great. thanks
L0K1T0
@l0k1t07_twitter
Hello everyone! Just wanted to know if I can use it to replicate an old DBF database to another.
joshahicks
@joshahicks
You can replicate from a DBF as a source but it would need to replicate into a relational database as the target
L0K1T0
@l0k1t07_twitter
Perfect!
That's what I need
I'll begin reading of how to implement this. Thanks
joshahicks
@joshahicks
There is a dbf router type
L0K1T0
@l0k1t07_twitter
I'll try to add some index on the target Relational DB.
Great! Thanks Josh
Its the last router type listed
You use a file trigger to monitor changes to your dbf file and then link this file trigger to a dbf router
L0K1T0
@l0k1t07_twitter
Thanks!!!
Peixoto258
@Peixoto258
Hello guys, I'm with a problem in symmetric 3.11
Since early the symmetric is with this message and don't going ahead.
image.png
The initial load is true and the inital load select is like 0 = 0
Someone can help me here? :D
Peixoto258
@Peixoto258
Hi guys, :D
How can I do to initial load to be faster?
I read to put bulk in data_loader_type, but it's the same thing
Is there any configurations? My source table has 2k records to make a initial load :/
joshahicks
@joshahicks
@Peixoto258 you may want to upgrade to latest 3.11.9 or even 3.12 I believe there was an issue around initial loads getting stuck
Peixoto258
@Peixoto258
Thanls @joshahicks