Markku.Savela@tel.vtt.fi (Markku Savela) (04/13/90)
I seemed to have a run-away /etc/init on my lone dn3500 (using TCP/IP, no other apollo domain nodes). When looked with "ps aux" etc/init had SZ ~ 20000 and RSS kept yoyoing from around 100 upto 2000-4000! You can guess what this does to 8Mb machine... (SZ was also slowly creeping upwards...). Any ideas what might be the problem and how to prevent it happening? I did have one socket buffer overlow (netstat showed sendq > 9000 for a while hours ago) and disk is 90% full, if those can have anything to do with it...? After reboot I hasn't repeated yet... -- Markku Savela | savela@tel.vtt.fi Technical Research Centre of Finland | Telecommunications Laboratory | Markku.Savela@vtt.fi Otakaari 7 B, SF-02150 ESPOO, Finland | savela%vtttel@router.funet.fi
ross@cancol.oz (Ross Johnson) (04/14/90)
In article <4815@hemuli.tik.vtt.fi>, Markku.Savela@tel.vtt.fi (Markku Savela) writes: > > I seemed to have a run-away /etc/init on my lone dn3500 (using > TCP/IP, no other apollo domain nodes). When looked with "ps aux" > etc/init had SZ ~ 20000 and RSS kept yoyoing from around 100 upto > 2000-4000! You can guess what this does to 8Mb machine... (SZ was > also slowly creeping upwards...). > [ First of all, if anyone sees this article can they send me e-mail. I'd like to know if it gets any further than the phone plug on the wall. ] I had this exact problem just recently and it had me worried. It turned out to be lines in the /etc/ttys file which specified "none" for the command field and were also flagged "on". I had done this so that root could rlogin by adding lines for all the ttyp* devices and following the manual verbatim, ie. ttys(4) said the line should be flagged "on" and "secure". I've found I only need the "secure" flag (unless there's something I haven't discovered yet). Ross Johnson ross@cancol.oz.au