home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: comp.unix.ultrix
- Path: sparky!uunet!usc!sol.ctr.columbia.edu!avogadro.barnard.columbia.edu!shenkin
- From: shenkin@avogadro.barnard.columbia.edu (Peter S. Shenkin)
- Subject: Another question; was: Re: Two questions from Ultrix novice
- Organization: Dept. of Chem, Barnard College, Columbia U, New York
- References: <1992Jul20.150115.2701@ctr.columbia.edu> <790982@athena.lkg.dec.com>
- Message-ID: <1992Jul21.162216.25343@ctr.columbia.edu>
- Sender: news@ctr.columbia.edu (The Daily Lose)
- Date: Tue, 21 Jul 1992 16:22:16 GMT
- X-Posted-From: avogadro.barnard.columbia.edu
- X-Posted-Through: sol.ctr.columbia.edu
- Lines: 53
-
- In article <790982@athena.lkg.dec.com> mamros@athena.lkg.dec.com (Shawn Mamros) writes:
- >
- >shenkin@avogadro.barnard.columbia.edu (Peter S. Shenkin) writes:
- >>1. The "df" command is not properly displaying disk utilization. If I
- >> add large files, the command properly tells me that more kbytes are
- >> used and fewer are free; however, if I remove large files df exhibits
- >> no change....
- >>
- >> What's broken, and how can I fix it?
-
- >What's almost certainly happening here is that a process still has the
- >file open when you remove it.....
-
- This is almost certainly not the answer, though it was a reasonable guess.
- The sysadm told me that this machine had this problem chronically. To check
- it out, after a reboot I did the following:
-
- 1. look at df output
- 2. copy large file (actually, "cp -r dirname new_dirname")
- 3. look at df output -- /usr has grown
- 4. "/bin/rm -rf new_dirname" -- it's gone
- 5. look at df output -- /usr has not changed in size
- 6. reboot; look at df output -- /usr has shrunk to original size
-
- No processes were started to access the new files. But please read on....
-
- Now for the new question: Yesterday a DEC field engineer showed up to
- replace tghe mother board in all our DecStation 5000's. This action was
- initiated by DEC, and though the service guy didn't know exactly why
- they were doing this, as well as he knew, it was a combination of (1)
- added RF shielding to meet FCC regulations, and (2) firmware rev to
- fix some bugs that occur in some situations.
-
- Following the board replacement my df problem went away. However, a far
- more serious problem started: the machine slowed down incredibly. It
- seems to be thrashing all the time. To login on the console it takes 30
- seconds for session manager to appear, and over an additional minute for
- DECterms to appear, then more time for them to respond -- all this
- accompanied by continuous disk chatter.
-
- A set of about 12 Fortran files which, the previous day, had compiled in less
- than an hour, took far longer; after 1-1/2 hour it still had not finished
- with the second .f file. This was also accompanied by literally continuous
- disk chatter.
-
- Can anyone help with diagnosing/fixing this problem?
-
- -P.
- --
- ************************f*u*cn*rd*ths*u*cn*gt*a*gd*jb*************************
- Peter S. Shenkin, Department of Chemistry, Barnard College, New York, NY 10027
- (212)854-1418 shenkin@avogadro.barnard.columbia.edu shenkin@cunixf.BITNET
- ******************** The singular of "media" is "medium". ********************
-