Hi Henrik, I'm modifying the original filers2bb scripts to monitor disk/inode/snapshot/quotas data on netapp storage via rsh and all is working quite good. Disk space and Inode data are also saved to rrd and graphed by the hobbitd_rrd module as well as the qtree quotas. With the qtree quotas graphs I've a little problem caused I think from the fact that the shares are for windows and there are spaces inside the directory names.
Just to do some examples I create the same layout of the Disk spaces/Inodes used and i put directory after the volume names just to be in line with all the other disk statistics:
Filesystem 1k-blocks Used Available Use% Mounted On /vol/g_home/App_Overview 2.0G 120K 2.0G 0% /vol/g_home/App_Overview /vol/g_home/Applicativi Rete 2.0G 127M 1.9G 6% /vol/g_home/Applicativi Rete /vol/g_home/AttFisse 10G 8.3G 1.7G 83% /vol/g_home/AttFisse /vol/g_home/Audit2006DSI 2.0G 68M 1.9G 3% /vol/g_home/Audit2006DSI /vol/g_home/BO6.5 2.0G 732M 1.3G 36% /vol/g_home/BO6.5 /vol/g_home/Common 3.4G 3.1G 376M 89% /vol/g_home/Common
In the graph part I get problems with the line
/vol/g_home/Applicativi Rete 2.0G 127M 1.9G 6% /vol/g_home/Applicativi Rete
The first problem that I've solved was related to the first directory name (the left one) where with a " " in the name the hobbit engine will shift all data and do some odd graphing. I substituted the " " with a "_" in the first part and the graphs are now the correct ones but the directory names in the graph is wrong because it only get "Applicativi" in the name of the disk in the rrd. If possible I need to leave at least one of the 2 "Filesystem" or "mounted on" with a space inside because we also have some qtree with "_" and if I change all of them it will be more difficult to check where the problem is from people that don't know this behavious.
Henrik, it's possible that the disk graphing routine will get the entire last parameters until end of line so that with "Applicativi Rete" it will get the entire string or that the hobbit program will convert the " " before doing the rrd if the problem is in the rrd engine?
I've also another question, I'm getting data about snapshot directories to check the space occupied by them and it could be nice to have also the graph of those disk spaces. The format I'm getting is the same of the disk/inodes/qtree and the column is called snapshot.
This is a data sample of what I get: Filesystem 1k-blocks Used Available Use% Mounted On /vol/rootvol/.snapshot 4194304 137036 4057268 3% /vol/rootvol/.snapshot /vol/CVS/.snapshot 1048576 1320 1047256 0% /vol/CVS/.snapshot /vol/g_home/.snapshot 57671680 4596052 53075628 8% /vol/g_home/.snapshot /vol/n_home/.snapshot 5242880 234968 5007912 4% /vol/n_home/.snapshot
It's possible to put in the hobbit server that if it get the column snapshot it will try to save rrd files and graph them as for the other disks columns?
Thanks for your good work with hobbit ! It's really good and saving a lot of time :)
Francesco Duranti