Just to send out an update on this issue .... The Linux servers have 2 NIC connecting through two seperate switches in an Active/Active bonding. We changed the switch configuration to enable Flow Control and updated the Intel drivers for the NICs and switched them to Active/Passing. Then, we got much better performance -- a 100MB ArchiveLog written in 8-12 seconds. Also, the "corrupt block" error on the last ArchiveLog went away in repeated recovery tests. Hemant >Date: Tue, 07 Dec 2004 00:18:08 +0800 >To: oracle-l@xxxxxxxxxxxxx >From: Hemant K Chitale <hkchital@xxxxxxxxxxxxxx> >Subject: ArchiveLogs on NetApps ? > > > >We've just begun our first NetApps [9iRAC on Linux] implementation. > >What I've found in testing so far is > > a) Writing ArchiveLogs to a NetApps Filer volume seems to be very slow > [5min for a 100MB >logfile against 30seconds on local disks] Do the ArchiveLog Dest volumes >have to be mounted >with the same NFS options as the Database Files volumes ? > >b) Testing recoveries using SNAP CREATE and SNAP RESTORE we find that we >frequently >encounter a corrupt block in the last ArchiveLog being written to the >ArchiveLog Dest volume. >Any ideas why ? > >Backup : > BEGIN BACKUP > Snap Create Database Files Volume > END BACKUP ;; ARCHIVE LOG CURRENT a few times > Snap Create ArchiveLog Files Volume > >Recovery : > Snap Restore Database Files Volume > Snap Restore ArchiveLog Files Volume > RECOVER DATABASE USING BACKUP CONTROLFILE UNTIL CANCEL > > > > >Hemant K Chitale >http://web.singnet.com.sg/~hkchital Hemant K Chitale http://web.singnet.com.sg/~hkchital -- //www.freelists.org/webpage/oracle-l