While attending a webex for a SRM solution, couple of questions came up with regards to the number of files in file system and average file size per file system. I should write a snmp query for getting that info for Netapps. Most SRM/file reporting solutions’ performance depends on the number of files, not the size of the file system.
technologist defying the peter principle
Somewhat Related
- Who Am I ?
- prototype 1.5.0 patch for Safari
- Leopard and DFS
- ibrix user quotas
- utterly useless unix command (uuuc)
2 Comments
Share this post
On September 18, 2006 @ 7:51 pm, rajeev karamchedu said:
Thx for the note. We/I know about the filestats command. I even use it in a few occasions. However, that requires access to the console and not very helpful to users. At our location, non-admins do not have any console access. My next post attempts to provide a shell routine (about to become a php routine)
I also experienced high load on the filers when running filestats.
Netapp does have an exhaustive SNMP mib. I provide several php routines that are used to get the filer info as part of our workflow.
AJAX and Netapp MIB is a killer combination!
You can use filestats on the filer to get filesize information from a filer. It’s on a volume basis, so if you have a lot of volumes, I see a for loop in your future. The syntax looks like this.
filestats sizes * snapshot hourly.0 volume vol03
That will give you counters for every unique filesize in volume vol03, in snapshot hourly.0. Filestats only runs against a snapshot.
If you wanted more in buckets, you could do this.
filestats ages 1k 2k 3k 4k 5k 6k 7k 8k 9k 10k snapshot hourly.0 volume vol03
I warn you if you run this you will loose your console until the command is done running. This can take anywhere from a few minutes to a few hours (not suprisingly) based off the number of files you have. I therefor recommend running doing this via rsh.
You could graph this fairly easily I think, might be nice to have a running average or plot things over the lifespan of a filer.