3

I am getting the error "memory mapped file write failed" while trying to run Snap Shot Agent in Transactional Replication. Snap shot works fine if the table is having only a few hundreds of rows. And fails if the table contains few thousands of rows. The SQL Server is installed on a Linux Server and tried many suggestions found online. But to no avail. Can anyone please guide me on this.

Screen shot of Transaction Replication Error

Rahul Indian
  • 31
  • 2
  • Maybe check your hard disk for bad blocks and other failing conditions. – Rex Henderson Apr 01 '21 at 05:45
  • No bad blocks found. – Rahul Indian Apr 01 '21 at 08:58
  • I have encountered this issue too just now. Also running SQL Server in a container trying to configure transactional replication. I haven't found any help. We may have to go back to running SQL on Windows – KellySandwiches Jun 17 '21 at 13:28
  • A different thread in SO mentioned that it seemed to be a new problem with CU9 (they rolled back to CU8 and it worked). – mbourgon Jul 02 '21 at 20:48
  • I have an open case with MS on it. They've been able to reproduce it with AdventureWorks, and are currently still investigating. Possible workaround: roll back to CU8. – mbourgon Oct 07 '21 at 20:37

1 Answers1

0

We've also had a similar issue with Snapshot Replication. As a workaround, we split the replication into multiple publications. After several trials, we were able to successfully replicate our database which contains more than 1000 tables with 5 publications. Except for 1 table with a text column (containing big xml), maybe we need another workaround in this case.

If you are only replicating 1 large table, you might try splitting it using Filter Rows.

We are using SQL Server 2019 Developer Edition on Ubuntu 18.04 LTS Server Linux Container with 8GB RAM. Other databases are installed on this server: MySQL 8.0, MariaDB 10.5, and PostgreSQL 12.5.

ragazenta
  • 49
  • 3
  • Thanks for the reply. I am trying with a single table in a single publication. Even with filters applied there are cases with large number of rows in a table. – Rahul Indian May 12 '21 at 04:50