[Bucardo-general] NUMLOG / BUCARDO - Only half of insert are done in database
Support EDI
support.edi at numlog.fr
Fri Dec 2 10:46:43 UTC 2022
Hi,
Many thanks for your answer. I confirm that replication works fines when
I only insert one raw after another.
I have set up the solution provided by Francisco Puga and it works fine.
I really appreciate your help thank you very much for your availability.
Best regards,
*Quentin BIENFAIT *
Equipe EDI / EDI Team
*_Mail :_ support.edi at numlog.fr <mailto:support.edi at numlog.fr> *
*_Tél :_ +33 1 30 79 16 16 (choix 2) <tel:+33130791616> *
9 ter route de Saint Germain
78640 Villiers Saint Frédéric - France
linkedin numlog <https://www.linkedin.com/company/numlog/> linkedin
<https://www.linkedin.com/company/numlog/> formulaire newsletter numlog
<https://cce6ca04.sibforms.com/serve/MUIEAA5rng8bgYTlAybtSwmIjNm02i1D75jtRGGJKYmaDWA0p1F7pFksNkRD2XGHdmuAWCVoAkNLyYLSK0dJAV2FbkodU2DNptDAf4ITyjy4qK2nteGxkYQkUOM1dB65fTbAvmk7mAQkokRC02MAYzFPkVWDN_cbdU5dcNeHwpkaa2PAVFkeiSe8OQ05-QMsu4CUpRzSswD6zk_z>
newsletter
<https://cce6ca04.sibforms.com/serve/MUIEAA5rng8bgYTlAybtSwmIjNm02i1D75jtRGGJKYmaDWA0p1F7pFksNkRD2XGHdmuAWCVoAkNLyYLSK0dJAV2FbkodU2DNptDAf4ITyjy4qK2nteGxkYQkUOM1dB65fTbAvmk7mAQkokRC02MAYzFPkVWDN_cbdU5dcNeHwpkaa2PAVFkeiSe8OQ05-QMsu4CUpRzSswD6zk_z>
site web numlog <https://www.numlog.fr> numlog.fr <https://www.numlog.fr>
Ce message électronique contient des informations confidentielles,
couvertes par le secret professionnel ou réservées exclusivement à leur
destinataire. Toute lecture, utilisation, diffusion ou divulgation sans
autorisation expresse est strictement interdite.
Si vous n'en êtes pas le destinataire, merci de prendre contact avec
l'expéditeur et de détruire ce message.
This message is for the designated recipient only and may contain
privileged, proprietary, or otherwise private information. If you have
received it in error, please notify the sender immediately and delete
the original. Any other use of the email by you is prohibited.
Avant d'imprimer, pensez à l'environnement - Before printing, think
about the environment
Le 02/12/2022 à 08:41, Perez Sanchez, Enrique a écrit :
> Hi!
>
> Your setup looks good to me, Quentin.
>
> Before you try to insert data on both servers simultaneously your replication works ok, right?
>
> Please, check this excellent reply by Francisco Puga on how to replicate DBs with sequences on multi-master with Bucardo:
> https://stackoverflow.com/questions/30752829/how-to-handle-sequences-in-bucardo-postgresql-multi-master
>
> I guess you’re getting close now :)
>
> BR!
> Kike
>
> ----
> From: Support EDI<support.edi at numlog.fr>
> Sent: jueves, 1 de diciembre de 2022 18:52
> To: Perez Sanchez, Enrique (SMO SWE RC-ES EN EN ITSEC&PLT-COC PLT)<enrique.perez_sanchez at siemens.com>;bucardo-general at bucardo.org
> Subject: Re: [Bucardo-general] NUMLOG / BUCARDO - Only half of insert are done in database
>
> Hello Kike,
>
> Many thanks for your quick feedback and sorry by my late one.
>
> I have re installed Bucardo to follow your advice. I have now 1 instance of bucardo installed on server 1, nothing on server 2. Both databases are now set to source.
> I think now I got more details on the error. Now when I'm trying to insert several datas at the same time I got a primary key constraint violation on my second server.
> I have added sequences as I though it should help to solve it but I probably did something wrong.
>
> Below is my setup :
>
> -- dbgroups:
> dbgroup: hbmsync Members: serv1:source serv2:source
> -- databases:
> Database: serv1 Status: active Conn: psql -U bucardo -d hubmaster -h ***.***.***.24
> Database: serv2 Status: active Conn: psql -U bucardo -d hubmaster -h ***.***.***.24
> -- relgroup:
> Relgroup: hubmaster DB: serv1 Members: public.files
> Used in syncs: hbmsync
> -- syncs:
> Sync "hbmsync" Relgroup "hubmaster" DB group "hbmsync" serv1:source serv2:source [Active]
> -- tables:
> 2. Table: public.files DB: serv1 PK: code (integer)
> -- sequences:
> Sequence: public.files_code_seq DB: serv1
>
> I have also tried with a second sequence with db pointing to server2 but the error remain the same.
>
> Have you any idea on how I can set up bucardo to fix this issue?
>
> Many thanks for your help.
>
> Best regards,
>
>
>
> Quentin BIENFAIT
> Equipe EDI / EDI Team
>
>
> mailto:support.edi at numlog.fr
> tel:+33130791616
>
> 9 ter route de Saint Germain
> 78640 Villiers Saint Frédéric - France
>
>
>
> https://eur01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fwww.linkedin.com%2Fcompany%2Fnumlog%2F&data=05%7C01%7Cenrique.perez_sanchez%40siemens.com%7C57decff698c84d43b79e08dad3c4b4fd%7C38ae3bcd95794fd4addab42e1495d55a%7C1%7C0%7C638055139062663379%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&sdata=Lh2D1Cbs4vYt8NBTS9kNNTvo16QnuTc9dMmgOOJ3t0w%3D&reserved=0
>
> https://eur01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fcce6ca04.sibforms.com%2Fserve%2FMUIEAA5rng8bgYTlAybtSwmIjNm02i1D75jtRGGJKYmaDWA0p1F7pFksNkRD2XGHdmuAWCVoAkNLyYLSK0dJAV2FbkodU2DNptDAf4ITyjy4qK2nteGxkYQkUOM1dB65fTbAvmk7mAQkokRC02MAYzFPkVWDN_cbdU5dcNeHwpkaa2PAVFkeiSe8OQ05-QMsu4CUpRzSswD6zk_z&data=05%7C01%7Cenrique.perez_sanchez%40siemens.com%7C57decff698c84d43b79e08dad3c4b4fd%7C38ae3bcd95794fd4addab42e1495d55a%7C1%7C0%7C638055139062819600%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&sdata=R1xIuZrf4szglFXkAjUe5x5zAIeNvqvhBfOZAfz8Rm8%3D&reserved=0
>
> https://eur01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fwww.numlog.fr%2F&data=05%7C01%7Cenrique.perez_sanchez%40siemens.com%7C57decff698c84d43b79e08dad3c4b4fd%7C38ae3bcd95794fd4addab42e1495d55a%7C1%7C0%7C638055139062819600%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&sdata=7VPw7qOVFLBDoGzKxRVQz2zmbuq1%2FQfkXCJL6nvOTlA%3D&reserved=0
>
> Ce message électronique contient des informations confidentielles, couvertes par le secret professionnel ou réservées exclusivement à leur destinataire. Toute lecture, utilisation, diffusion ou divulgation sans autorisation expresse est strictement interdite.
> Si vous n'en êtes pas le destinataire, merci de prendre contact avec l'expéditeur et de détruire ce message.
> This message is for the designated recipient only and may contain privileged, proprietary, or otherwise private information. If you have received it in error, please notify the sender immediately and delete the original. Any other use of the email by you is prohibited.
> Avant d'imprimer, pensez à l'environnement - Before printing, think about the environment
>
>
> Le 25/11/2022 à 12:38, Perez Sanchez, Enrique a écrit :
> Bonjour, Quentin!
>
> Are you trying to set up 2 instances of Bucardo, one replicating from A to B and another one from B to A? (I think there may be a typo in the part where you paste your configuration, you said “server1” twice).
> Please, try to set up just 1 instance of Bucardo, and try adding both databases as “source”. Then please repeat the tests and let’s see how that works.
>
> When you create a new sync with 2 databases in 2 nodes (with a single instance of Bucardo), you will notice that a new schema (bucardo) appears on all the databases and some triggers appear on all tables.
> In this schema there are some tables that bucardo uses to keep track of the “deltas” as well as some tables with per-table (in the public or the other schemas) information.
> Why am I telling you this? Because this is key: You can have bidirectional replication between nodes A and B with Bucardo software installed only on node A, B, or even on a different node C!
> If Bucardo software is installed in node A and node A goes down, triggers in the tables in node B will populate the tables in the bucardo schema in node B. When node A goes back up, the Bucardo software processes (KID and CTL) will read and process those contents and process the replication, handling conflicts should there be any (“man bucardo” and search for “conflict”).
>
> Please, try and let us know. And @rest, please correct me if I’m wrong on anything :)
> Good luck!
>
> Kike
>
> From: Bucardo-generalmailto:bucardo-general-bounces at bucardo.org On Behalf Of Support EDI via Bucardo-general
> Sent: jueves, 24 de noviembre de 2022 11:51
> To:mailto:bucardo-general at bucardo.org
> Subject: [Bucardo-general] NUMLOG / BUCARDO - Only half of insert are done in database
>
> Dear Bucardo team,
>
> Few weeks ago I started working on building a postgresql cluster (in master / master mode) composed of 3 VM with Debian 11 installed.
> 1. 1 server is running an HA proxy server (version 2.2.9). It is used to provide load balancing features
> 2. 2 Postgresql server (version 13.8) and Bucardo (version 5.6.0). Bucardo is used here to provide high availability with a master / master mode.
> I have tested inserting data in each database server locally and every data is available on each server. So the master / master looks working fine.
> Then I have tried to insert data from a distant server through the HA Proxy server and I have noticed that only the half of the insert are available in postgresql.
>
> I have analyzed logs of the 3 softwares to find where something goes wrong and think it comes from bucardo.
> In this case I tried to insert 4 rows in a table named "files". In postgresql only 2 entries are available. The 2 entries are the same on both postgresql server.
> 1. First I checked HA proxy logs and find that it was well redirecting 50% of insert on server 1 and 50% on server 2. No error in its logs so I presume that it really works fine.
> 2. Then I checked Postgresql log. In Postgresql logs I cannot see any error and I can't see any attempt to insert the missing values. I have changed log level to ensure I'm not missing something but without success.
> 3. At last I have checked in Bucardo and find below logs :
> (292750) [Mon Nov 21 10:46:30 2022] KID (sync_oneToTwo) Got NOTICE run_sync_oneToTwo from 292748 (Bucardo DB) (line 3113)
> (292750) [Mon Nov 21 10:46:30 2022] KID (sync_oneToTwo) Adding entry to syncrun table
> (292750) [Mon Nov 21 10:46:30 2022] KID (sync_oneToTwo) Populating the dbrun table
> (292750) [Mon Nov 21 10:46:30 2022] KID (sync_oneToTwo) Adding note to the syncrun table
> (292750) [Mon Nov 21 10:46:30 2022] KID (sync_oneToTwo) Doing final maindbh commit
> (292750) [Mon Nov 21 10:46:30 2022] KID (sync_oneToTwo) Set database "serv1" to repeatable read read write
> (292750) [Mon Nov 21 10:46:30 2022] KID (sync_oneToTwo) Set database "serv2" to repeatable read read write
> (292750) [Mon Nov 21 10:46:30 2022] KID (sync_oneToTwo) Checking truncate_trigger table on database "serv1"
> (292750) [Mon Nov 21 10:46:30 2022] KID (sync_oneToTwo) Sequence public.files_code_seq from db serv1 is the highest
> (292750) [Mon Nov 21 10:46:30 2022] KID (sync_oneToTwo) Tables with deltas on serv1: 1 Without: 0
> (292750) [Mon Nov 21 10:46:30 2022] KID (sync_oneToTwo) Delta count for serv1.public.files : 2
> (292750) [Mon Nov 21 10:46:30 2022] KID (sync_oneToTwo) Total delta count: 2
> (292750) [Mon Nov 21 10:46:30 2022] KID (sync_oneToTwo) Setting session_replication_role to replica for database serv2
> (292750) [Mon Nov 21 10:46:30 2022] KID (sync_oneToTwo) Rows to push from serv1.public.files: 2
> (292750) [Mon Nov 21 10:46:30 2022] KID (sync_oneToTwo) Deleting from target serv2.public.files (round 1 of 1)
> (292750) [Mon Nov 21 10:46:30 2022] KID (sync_oneToTwo) Rows deleted from serv2.public.files: 2
> (292750) [Mon Nov 21 10:46:30 2022] KID (sync_oneToTwo) Copying from serv1.public.files
> (292750) [Mon Nov 21 10:46:30 2022] KID (sync_oneToTwo) Rows copied to (postgres) serv2.public.files: 2
> (292750) [Mon Nov 21 10:46:30 2022] KID (sync_oneToTwo) Totals: deletes=2 inserts=2 conflicts=0
> (292750) [Mon Nov 21 10:46:30 2022] KID (sync_oneToTwo) Rows inserted to bucardo_track for serv1.public.files : 2
> (292750) [Mon Nov 21 10:46:30 2022] KID (sync_oneToTwo) Setting session_replication_role to default for database serv2
> (292750) [Mon Nov 21 10:46:30 2022] KID (sync_oneToTwo) Issuing final commit for all databases
> (292750) [Mon Nov 21 10:46:30 2022] KID (sync_oneToTwo) All databases committed
> (292750) [Mon Nov 21 10:46:30 2022] KID (sync_oneToTwo) Total time for sync "sync_oneToTwo" (2 rows, 1 table): 0.05 secondes
> (292750) [Mon Nov 21 10:46:30 2022] KID (sync_oneToTwo) Sending NOTIFY "ctl_syncdone_sync_oneToTwo" (line 4937) skip_commit=0
> I'm not sure I correctly understand bucardo logs as I'm not experimented using it, but it looks like it deletes 2 raws in a server 2 and has pushed 2 raws from server 1.
> My first question is, is my analysis correct or am I wrong?
> My second question if I'm right is how should I set up Bucardo to avoid this issue. I'm sure I'm doing something wrong but I can't find what's wrong.
> My second question if I'm wrong, do you have any idea of the cause of the issue?
>
> On server1, my bucardo set up is :
>
> -- dbgroups:
> dbgroup: sync_oneToTwo Members: serv1:source serv2:target
> -- databases:
> Database: serv1 Status: active Conn: psql -p 5432 -U bucardo -d hubmaster -h ***.**.**.24
> Database: serv2 Status: active Conn: psql -p 5432 -U bucardo -d hubmaster -h ***.**.**.25
> -- relgroup:
> Relgroup: one_two DB: serv1 Members: public.files, public.files_code_seq
> Used in syncs: sync_oneToTwo
> -- syncs:
> Sync "sync_oneToTwo" Relgroup "one_two" DB group "sync_oneToTwo" serv1:source serv2:target [Active]
> -- tables:
> 2. Table: public.files DB: serv2 PK: code (integer)
> 1. Table: public.files DB: serv1 PK: code (integer)
> -- sequences:
> Sequence: public.files_code_seq DB: serv1
> Sequence: public.files_code_seq DB: serv2
>
> On server1, my bucardo set up is :
>
> -- dbgroups:
> dbgroup: sync_oneToTwo Members: serv1:source serv2:target
> -- databases:
> Database: serv1 Status: active Conn: psql -U bucardo -d hubmaster -h ***.**.**.25
> Database: serv2 Status: active Conn: psql -U bucardo -d hubmaster -h ***.**.**.24
> -- relgroup:
> Relgroup: one_two DB: serv1 Members: public.files, public.files_code_seq
> Used in syncs: sync_oneToTwo
> -- syncs:
> Sync "sync_oneToTwo" Relgroup "one_two" DB group "sync_oneToTwo" serv1:source serv2:target [Active]
> -- tables:
> 1. Table: public.files DB: serv1 PK: code (integer)
> 2. Table: public.files DB: serv2 PK: code (integer)
> -- sequences:
> Sequence: public.files_code_seq DB: serv1
> Sequence: public.files_code_seq DB: serv2
>
> Thanks in advance for any help provided.
>
> Best regards
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://bucardo.org/pipermail/bucardo-general/attachments/20221202/41e93b8f/attachment-0001.htm>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: BVlKnHEX1rWPiOX0.png
Type: image/png
Size: 280656 bytes
Desc: not available
URL: <https://bucardo.org/pipermail/bucardo-general/attachments/20221202/41e93b8f/attachment-0005.png>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: HbbYy59L9cGUsDon.png
Type: image/png
Size: 6472 bytes
Desc: not available
URL: <https://bucardo.org/pipermail/bucardo-general/attachments/20221202/41e93b8f/attachment-0006.png>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: Z0bhgdQVGP8bKxOh.png
Type: image/png
Size: 16561 bytes
Desc: not available
URL: <https://bucardo.org/pipermail/bucardo-general/attachments/20221202/41e93b8f/attachment-0007.png>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: PgJgus1YyU6t6Ptj.png
Type: image/png
Size: 16561 bytes
Desc: not available
URL: <https://bucardo.org/pipermail/bucardo-general/attachments/20221202/41e93b8f/attachment-0008.png>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: cSpQXLMXpRB4mSMo.png
Type: image/png
Size: 16561 bytes
Desc: not available
URL: <https://bucardo.org/pipermail/bucardo-general/attachments/20221202/41e93b8f/attachment-0009.png>
More information about the Bucardo-general
mailing list