home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!destroyer!ncar!hsdndev!news.cs.umb.edu!rouilj
- From: rouilj@cs.umb.edu (John P. Rouillard)
- Newsgroups: comp.unix.admin
- Subject: Re: does anyone have a disk x% full capture program???
- Message-ID: <ROUILJ.92Nov17130357@ra.cs.umb.edu>
- Date: 17 Nov 92 18:03:57 GMT
- References: <fish.721343514@news2.gsfc.nasa.gov> <meskes.721482983@ulysses>
- Sender: news@cs.umb.edu (USENET News System)
- Organization: University of Massachusetts at Boston
- Lines: 37
- In-Reply-To: meskes@ulysses.informatik.rwth-aachen.de's message of 11 Nov 92 11: 56:23 GMT
- Nntp-Posting-Host: ra.cs.umb.edu
-
-
- Has anybody enhanced watcher to support LINE selection
- I.E. I would like something like (for a df command):
-
- LINE 1 header1%k
- LINE 2 header2%k
- LINE 3-0 1 disk%k 2 3 kbytes_used%d 4 kbytes_free%d 5 percent_used%d
-
- for the following df output:
-
- Filesystem Total kbytes kbytes %
- node kbytes used free used Mounted on
- /dev/rz6h 494294 412044 32821 93% /nfs/news
-
- Line 1 and line 2 would be header files and would be reported in the
- watcher output. Lines 3 and on would be analyzed according to the
- template given. This would allow the removal of the tails and greps
- that I have to use now to filter the output of the commands.
-
- Also allowing expressions in the spec of things to check would be very
- useful as well. Given the output of nfstat -c:
-
- Client rpc:
- calls badcalls retrans badxid timeout wait newcred
- 4340 1 84 2 84 0 0
-
- LINE 3 1 calls%d 2 badcalls%d 3 retrans%d 4 badxid%d 5 timeout%d :
-
- retrans/calls*100 0 1 # the percentage of retrans < 1%
- badcalls/retrans*100 0 1 # same idea.
-
- Right now all of these numbers are being calculated using mawk, and
- perl. Needless to say this is causing the watcher program and the
- programs it runs to be influencing the load and other operations on
- the machine it is watching.
-
- -- John
-