UNIX for Dummies Questions & Answers

Talk about making me feel welcome. Is nice!

OK, to the question. I work in the IT dept for a Financial Investment Management company. I've written a VB script to check various systems and the script works well but I've been asked to add an additional check on a UNIX system (specifically a NetApp SAN which is used from our VMWare ESX environment) which is where this forum comes in to the rescue.

I've established the 'df -h' command will produce the output I need, albeit with some extra columns and some extra lines, but I need to work out how to get the data somewhere from which our Windows clients can see it and on a schedule.

So
Just to clarify point 1, here's an example of the output:

Code:

Filesystem     total    used   avail   capacity     Mounted on
/vol/TestVol0/ 239GB   216GB    23GB        90%  /vol/TestVol0/
/vol/TestVol0/.snapshot 0MB 8023MB 0MB ---% /vol/TestVol0/.snapshot
/vol/vol0/      16GB   607MB    15GB         4%  /vol/vol0/
/vol/vol0/.snapshot 4096MB 11MB 4084MB 0% /vol/vol0/.snapshot

...which just looks a mess. It would look much neater (and be easier to query) if it was like this:

Code:

Filesystem      total    used   avail   capacity       Mounted on
/vol/TestVol0/  239GB   216GB    23GB        90%   /vol/TestVol0/
/vol/vol0/       16GB   607MB    15GB         4%       /vol/vol0/

...but not essential (note that doubled-up spaces have been stripped out by this interface so it still doesn't look as neat as it should here. Oh well).

TIA!

JJ

Depends which OS your running, but:

1) grep -v (ie: df -h | grep -v snapshot) - do a "man grep" - perhaps redirect output to a file
2) Take a look at the man pages for cron
3) easy way would be to ftp the file to the local network - you could also cron an ftp process to transfer the file for you at a regular time each day

Thank you.

I've found that the command grep isn't actually recognised, which is odd.

Is there a command you can run to establish exactly which OS is running?

JJ

you can try "uname -a"... but if you are on a netapp filer it should be some kind of BSD derivat?!

the rightmost column of 'uname -a' usually gives it up for the OS. what shell are you running under though? grep should be available on most systems...and shells.

Otherwise, check out one of the UWIN or Cygwin apps which would let you run the whole shebang from within Windows. Both even provide for ssh clients that can tap into your UNIX system's ssh service as well. It provides a full ksh93 option for those of us cursed with bills and gates...

Good call. 'uname' is also not recognised. When I log on with telnet it tells me this (about to try the CODE tags, thanks for that in the earlier post, BTW)

Code:

Data ONTAP

Not sure where to go from here as it seems to me if the system doesn't recognise basic commands like uname and grep what chance have I got of setting it up to FTP or working with CRON to schedule anything?

I should say that so far NetApp support haven't exactly showered themselves in glory hence I've come here and already got further!

i think the netapp os (ONTAP) is stripped from some unix basic functions... so it would be best to do the commands from remote with "ssh" which i think should be onboard with ONTAP. just use you admin workstation (which is hopefully not windows!) and work with ssh commands like this:

Code:

ssh -l username df -h | grep -v snapshot

if this works, we can look on to solve the other problems...

btw. if you run windows on your admin station you can maybe download a virtual desktop system like virtualbox and run a linux or unix on your workstation. this helps in two ways:
you can try commands on your workstation
and
you can use the tools from unix/linux to work on the output of the netapp mashine

hth,
DN2