Do re-attempts start over or continue where backup left off

jlynnmc10

I am getting the following errors when performing a full database backup. From these errors it appears that RedGate is re-attempting to back up the database. On a re-attempt does the write process start over or are the writes advancing, meaning when it preforms re-attempt 1 does it start all the way over writing the backup or does it pick up where it left off?

Job 'DBA - Full Backup (All Databases)' : Step 1, 'Backup (TSQL)' : Began Executing 2011-08-10 17:18:44

SQL Backup v6.5.1.9
Backing up ANAQUADOC_PRD (full database) on A instance to:
\\bubbly\sql_Anaqua\IAMPRDDB01_A\ANAQUADOC_PRD\ANAQUADOC_PRD_SB_20110810173633.BAK

Warning 210: Thread 0 warning:
WriteFile failed for file: \\bubbly\sql_Anaqua\IAMPRDDB01_A\ANAQUADOC_PRD\ANAQUADOC_PRD_SB_20110810173633.BAK at position: 256024511488
8/10/2011 8:01:05 PM: WriteFile failed for file: \\bubbly\sql_Anaqua\IAMPRDDB01_A\ANAQUADOC_PRD\ANAQUADOC_PRD_SB_20110810173633.BAK (The specified network name is no longer available.)
8/10/2011 8:01:05 PM: CloseTargetFile.FlushFileBuffers error: The specified network name is no longer available.

8/10/2011 8:01:35 PM: Re-attempt: 1
8/10/2011 8:02:44 PM: File (\\bubbly\sql_Anaqua\IAMPRDDB01_A\ANAQUADOC_PRD\ANAQUADOC_PRD_SB_20110810173633.BAK) reopened and write succeeded (2097152 bytes) at position (256024511488).
WriteFile failed for file: \\bubbly\sql_Anaqua\IAMPRDDB01_A\ANAQUADOC_PRD\ANAQUADOC_PRD_SB_20110810173633.BAK at position: 257865810944
8/10/2011 8:06:40 PM: WriteFile failed for file: \\bubbly\sql_Anaqua\IAMPRDDB01_A\ANAQUADOC_PRD\ANAQUADOC_PRD_SB_20110810173633.BAK (The specified network name is no longer available.)
8/10/2011 8:06:40 PM: CloseTargetFile.FlushFileBuffers error: The specified network name is no longer available.

8/10/2011 8:07:10 PM: Re-attempt: 2
8/10/2011 8:08:02 PM: File (\\bubbly\sql_Anaqua\IAMPRDDB01_A\ANAQUADOC_PRD\ANAQUADOC_PRD_SB_20110810173633.BAK) reopened and write succeeded (2097152 bytes) at position (257865810944).
WriteFile failed for file: \\bubbly\sql_Anaqua\IAMPRDDB01_A\ANAQUADOC_PRD\ANAQUADOC_PRD_SB_20110810173633.BAK at position: 372282229760
8/11/2011 12:31:35 AM: WriteFile failed for file: \\bubbly\sql_Anaqua\IAMPRDDB01_A\ANAQUADOC_PRD\ANAQUADOC_PRD_SB_20110810173633.BAK (The specified network name is no longer available.)
8/11/2011 12:31:35 AM: CloseTargetFile.FlushFileBuffers error: The specified network name is no longer available.

petey

It picks up where it last left off. It can't start from the beginning because it never stored the earlier backup data that has been written. That would take up way too much resources for large backups.

jlynnmc10

We are having severe performance issues with our network share which is being researched by the Network team but most of our large database retry 9 times and still never succeed which is only creating more contention on the network share. Is there a way to turn of the retries so I can do some testing? Thanks.

petey

Use the DISKRETRYCOUNT option e.g.

EXEC master..sqlbackup '-sql "BACKUP ... WITH DISKRETRYCOUNT = 0"'

jlynnmc10

Can I set DISKRETRYCOUNT to 0 in the registry? I would like to turn this off for one of our servers with a large database as the retries are frequently not successful and cause the backup to take 3-7 hours longer to fail than if successful. Thanks.

petey

No, it isn't possible to set the DISKRETRYCOUNT value in the registry to be used as the default value.