Hi, Yeah, by default mysqldump buffers the result of the "SELECT * FROM table" query in memory before writing the SQL statements (using mysql_store_result()). If you use the --opt option (or at least -q or --quick), it dumps the data as it gets it (using mysql_use_result()).
Hope that helps. Matt ----- Original Message ----- From: <[EMAIL PROTECTED]> Sent: Thursday, February 19, 2004 1:23 PM Subject: mysqldump via tcp/ip memory problem > > I've dumped alot of databases before using mysqldump, and am trying to > dump a larger database than normal, about 2.2GB in size.. The largest > table just over 12 million rows... It's dumping over a network to a > tape backup server.. > > I start the job off: > > /usr/local/bin/mysqldump -c -F --host=prv-master1 \ > --password=blahblah --port=3306 --user=blahblah --verbose mdb1 > > /tapesource/MDB1/mdb1.db > > It runs for bit, dumping some smaller tables, then gets the the largest > table (12mil row) .. runs for a bit and reports "Killed" > > Dmesg shows: > > __alloc_pages: 0-order allocation failed (gfp=0x1d2/0) > VM: killing process mysqldump > > Which leads to a memory problem, or lack of... The box does have approx. > 500MB of free ram... > > Is it just eating it up buffering the network response from the server? > > Mysqldump on client is Ver 8.22 Distrib 3.23.57 > Mysqld on server is 3.23.55-log > > Thoughts? -- MySQL General Mailing List For list archives: http://lists.mysql.com/mysql To unsubscribe: http://lists.mysql.com/[EMAIL PROTECTED]
