The SQL server does backups with this command to a network share \BackupServer\weekeklybackups"

BACKUP DATABASE DBNAME TO DISK = '\\BackupServer\weekeklybackups\DBNAME.BAK'

There is a 1 gig network between the SQL Databse server and the BackupServer. but it uses maximum 25% of the network, so the backup is taking for ever. If I make the database locally and copy it over to the same network share, it is faster using 99% of the network bandwidh.

So, the weekly automated backups are taking a long time, which is an issues

When I check the monitor it says it is waiting with "ASYNC_IO_...". When I did some google search it says, this is an issue on tape backup. But this is on iSCSI hard

disks. and manual copy over works fine. So, what could be the issue? It was working fine for a year until last week. Not sure what happened. but where should I be

looking in to?

Thanks for your time

closed as off-topic by Aaron Bertrand Sep 6 '14 at 12:19

If this question can be reworded to fit the rules in the help center, please edit the question.

  • 1
    Backup operation reads data from disk and writes to disk/tape. Now in your case I guess slowness is from SQL server side. May be the disk on which SQL server data files reside is facing |I/O contention which is causing whole process to slow down. Async IO completion is pointer to slow disk or I/O contention. Its quite possible because of SQL server under I/O pressure. Can you confirm load on server ? – Shanky Jul 14 '14 at 14:48
  • It would not affect data file as such but backup has to read data from disk where data files reside if disk is slow or facing I/O contention backup would not get I/O eaisly and would have to wait thus causing delay – Shanky Jul 14 '14 at 15:07
  • Thanks Shanky. but wouldn't that cause the slowness when I have the .bak file on the local hard drive where the data files are to the network share. It is faster in that. using 99% of the network link. But I will check the load on the server. – user3618129 Jul 14 '14 at 15:08
  • Shanky - screencast.com/t/irmCL3a01w. Here is the Activity monitor output. It is active but nothing alarming here. Should I check any other parameters for the load? – user3618129 Jul 14 '14 at 15:13
  • Can you post output for below counters for Object Physical disk: Avg. Disk sec/Transfer Avg. Disk Queue Length Avg. Disk sec/Read Disk reads/sec please measure this for more than 15 mins and post the value here. Please measure this counter for disk on which data file is residing – Shanky Jul 14 '14 at 15:20

It was an issues on the hard drive. It never showed any errors in event viewer or the controller. When we rebooted the server, the errors showed up and we replaced the drives.

Not the answer you're looking for? Browse other questions tagged or ask your own question.