Greetings,

I spent about 3+ days trying to get ASE & NetApp Filer to work properly. The platform I used was a Dual PIII 450 [1GB RAM] running Linux [RedHat 6.2] with 2.2.14 kernel and ASE 11.9.2.2 ESD #2 [EBF #8944]. The system has 2 Intel EtherExpress 10/100 cards. The NetApp Filer is a F720. I setup a direct network connection between the Linux server and the NetApp Filer [running 100 Base-T w/Full Duplex].

First of all I have to say the instructions located at http://www.netapp.com/tech_library/3040.html were only tested on Solaris and not any other flavor or UNIX. Some of the NFS options were different or did not exist under Linux so it took me a while to get that done. Basically you need to have the "lockd" [NFS Lock Daemon] running and also use the "nolock" option for the NFS mount in order to make it work. After that I learned that you can only create the devices under the mount point [i.e. if your mount point is "/sybase" your devices must be "/sybase/<device>] or it will fail. Again none of this was located in the documentation.

The creation of the server and devices was a lot better than creating file(s) on a RAID filesystem but about the same with using "partitions" on a RAID drive [with no filesystem]. Since ASE uses O_SYNC option [wait until the O.S. have verified your write before going forward] for filesystem files it ensure [to a point] that corruption of the database is minimized. However Sybase says that using filesystem files is still not a solution until ASE 12.0 [read below].

I had a working production database using this server and decided to do a little experiment. I did a backup of my database and then did a "reboot" of my system without shutting down Sybase. When the system came back up I tried to restart ASE but found out that the "sysusages" information for my database was corrupted. Working with Sybase I tried everything possible but had to drop/re-create my database from scratch. I called and informed NetApp of my problems and asked them to investigate it. It has been over 7 business days and I have not heard any further information from them.

All in all I do not believe that using filesystem files is a solution for ASE on any platform. With ASE 12.0 they now switch to D_SYNC [wait until the write has been comitted to the file before continuing on] for filesystem files. This of course will make using filesystem files more of an option *but* it will also slow things down [you won't take the OS at it's word anymore] since you will guarantee that data will be written to your filesystem devices.

Until NetApp can prove to me that their solution is faster than locally attached RAID using "partitions"/"raw I/O" [raw I/O is part of Linux 2.4.X kernel and probably ASE 12.5 for Linux] where we can do asynchronous read/write [filesystem will *always* be synchronous] I would not recommend it to anyone.

Yours,

Shahryar