Unsupported Screen Size: The viewport size is too small for the theme to render properly.

"copy Table To Different Host/database…" Issue

forums forums SQLyog SQLyog: Bugs / Feature Requests "copy Table To Different Host/database…" Issue

  • This topic is empty.
Viewing 0 reply threads
  • Author
    Posts
    • #11705
      vygi
      Member

      Greetings,

      we've got following problem with the SQLyog 8.15 beta1:

      “Copy Table to Different Host/Database…” is extremely slow now for bigger tables when “Use Bulk INSERT” is activated!

      Right now I'm trying to copy a table with ca. 18000 rows (slightly more than 1 MB of data in total) from one MySQL 5.1.39 server to another.

      It already takes more than 3 minutes. SHOW PROCESSLIST on the source server reports “Writing to net” for 200+ seconds. The query is “select * from source_table” without any limits.

      Whan after 1-2 minutes of waiting I click on “Stop”, it still takes another 3 to 6 (!) minutes to stop (“Aborting copy…”). SHOW PROCESSLIST on the source server shows another “select * from source_table” during the first 2-4 minutes, and no activities at all for the following 1-2 min.

      Same query (“select * from source_table”) returns 17628 rows in 0.06 sec when executed manually in Linux shell. Indeed it takes 20-30 seconds to display data on the screen but it is still much faster than SQLyog.

      Without “Use Bulk INSERT”, the copy takes around 10 seconds.

      max_allowed_packet is actually 32M on both servers.

      Same problem when trying to copy the table from one database within one server, from one database to another: quite fast without “Bulk Insert” but does not work (anymore) with it!

      Small tables (under 1000 rows) can be copied fast even with “Bulk Insert”.

      ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

      It was all fine earlier, when we were using SQLyog 8.13 and MySQL 5.0 on one of the servers.

      EDIT: table size was incorrect specified: barely more than 1 MB and not 110 MB!

Viewing 0 reply threads
  • You must be logged in to reply to this topic.