home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!mcsun!Germany.EU.net!kommu!norisc!georg
- From: georg@norisc.erls01.siemens.de (Georg Biehler)
- Newsgroups: comp.unix.sys5.r4
- Subject: Re: Dell - rmfree map overflow???
- Message-ID: <1140@kommu.UUCP>
- Date: 23 Nov 92 06:55:33 GMT
- References: <18NOV199209143914@apsicc.aps.edu>
- Sender: news@kommu.UUCP
- Lines: 47
- X-Newsreader: TIN [version 1.1 PL7]
-
- E. Drew Einhorn (einhorn_d@apsicc.aps.edu) wrote:
- [...]
- : I am getting a large number of messages of the form:
- :
- : WARNING: rmfree map overflow D01B475C. lost 1 items at -19yxxx
- :
- : where y is 0, 1, or 2 and x is a decimal digit. These messages
- : appear to be written to /dev/console not stdout or stderr. I
- : don't see them when running under X but when I exit X, I find
- : the last couple that haven't rolled off the top of the screen.
- :
- : Under Dell Unix SVR4 error messages don't accumulate in
- : /usr/adm/log. Is there a way to reconfigure so the
- : system error messages are logged.
-
- This messages get logged by the cmn_err() routine. They are accessible
- via /dev/osm, and get kept in /etc/.osm. After booting the system you
- find the old ones in /etc/.osm.old
- :
- : At the time the errors occure I am running a script to apply
- : patch "2" to Oracle V6.0.34.2.1 Oracle is attempting to
- : replace files in large .a archives. After getting many of these
- : errors the system eventually crashes and I have to hit the
- : power switch to get it restarted. When I try again after the
- : reboot and fsck's the archives are corrupted. Will pursue
- : this with Oracle and Dell Tech support in the morning. Hope
- : somebody recognizes this error message and I don't have to
- : resort to a horrendous find command looking for the string
- : "rmfree map overflow"
-
- From the "System Administrator's Guide (C-10)":
-
- rmfree map overflow # Lost # items
- Ran out of entries used to manage the available
- memory for the indicated map. The map overflow #
- is the address of the table that is too small; the
- commands "nm -x /stand/unix | grep #" can be used to
- get the name of the map table that overflowed. If the
- warnings persist, reconfigure the system with a larger
- table for the indicated map (The system has three map
- tables, whose sizes are determined by the SPTMAP,
- MSGMAP and SEGMAP tunables).
-
- -Georg
- ----------------------------------------------------------------------
- Georg H. Biehler E-Mail: georg@norisc.erls01.siemens.de
- RealTime SVR4 -> Try SORIX 386/486
-