michaelb@wshb.csms.com ( WSHB Operations Eng) (12/15/90)
I need some advice on kernal tuning for SCO Xenix 2.3.2 on a 20 MHz machine with 9 MB memory. Until recently my machine has been most heavily loaded during the day with only a few uucp connections and a login or two happening at night. Now - as the night crews are becoming more familiar with the machine, we start processing more reports with cron and all three modems fire up - I have started running into the maximum process limit. I don't really think there is a big problem of running out of memory. My swap space is huge and the 9 MB of ram should handle bunches of stuff anyway. I have looked at /usr/sys/conf/master and found that procs is set to 60. As I have never had to tune a kernal before I am curious what will happen if I just raise this limit to 120 and re-link. Are there any other things I should change at the same time to prevent bad things. I have included the tunable parameters list. (Does anyone have recomendations for a good book which explains what all of this stuff does?) ==================== * * The following entries form the tunable parameter table. * buffers NBUF 0 sabufs NSABUF 0 npbuf NPBUF 10 hashbuf NHBUF 128 dmaexcl DMAEXCL 1 inodes NINODE 100 files NFILE 100 mounts NMOUNT 8 calls NCALL 70 procs NPROC 60 clists NCLIST 100 locks NFLOCKS 50 shdata NSDSEGS 25 sdslots NSDSLOTS 3 maxproc MAXUPRC 30 memlim MEMLIM 100 swplim SWPLIM 30 timezone TIMEZONE (8*60) pages NCOREL 0 daylight DSTFLAG 1 cmask CMASK 0 maxprocmem MAXMEM 0 maxbuf MAXBUF 600 screens NSCRN 0 emaps NEMAP 10 shlsess NSXT 1 disks NDISK 2 nodename NODE "" msgmap MSGMAP (MSGSEG/2+1) msgmax MSGMAX (MSGSEG*MSGSSZ) msgmnb MSGMNB (MSGSEG*MSGSSZ) msgmni MSGMNI 10 msgtql MSGTQL 60 msgssz MSGSSZ 8 msgseg MSGSEG 1024 semmap SEMMAP (SEMMNS/2+1) semmni SEMMNI 10 semmnu SEMMNU 20 semmsl SEMMSL 10 semopm SEMOPM 5 semume SEMUME 5 semvmx SEMVMX 32766 semaem SEMAEM 16384 semmns SEMMNS 40 shmmax SHMMAX (1024*4096) shmmin SHMMIN 1 shmmni SHMMNI 25 shmseg SHMSEG 6 shmbrk SHMBRK 0 shmall SHMALL 4096 nqueue NQUEUE 1 nstream NSTREAM 1 nblk8192 NBLK8192 0 nblk4096 NBLK4096 0 nblk2048 NBLK2048 0 nblk1024 NBLK1024 0 nblk512 NBLK512 0 nblk256 NBLK256 0 nblk128 NBLK128 0 nblk64 NBLK64 0 nblk32 NBLK32 0 nblk16 NBLK16 0 nblk4 NBLK4 1 nblk1 NBLK1 NBLK8192+NBLK4096+NBLK2048+NBLK1024+NBLK512+NBLK256 nblk0 NBLK0 NBLK128+NBLK64+NBLK32+NBLK16+NBLK4 ndblock NDBLOCK NBLK0+NBLK1 nmblock NMBLOCK NDBLOCK nmuxlink NMUXLINK 1 nstrpush NSTRPUSH 0 nstrevent NSTREVENT 1 maxsepgcnt MAXSEPGCNT 0 strmsgsz STRMSGSZ 0 strctlsz STRCTLSZ 0 strlofrac STRLOFRAC 0 strmedfrac STRMEDFRAC 0 evqueues EVQUEUES 8 evdevs EVDEVS 16 evdevsperq EVDEVSPERQ 3 scrnmem SCRNMEM 0 kbtype KBTYPE 0 $$$ ========================= Thanks for any help. Michael -- Michael Batchelor--Systems/Operations Engineer #compliments and complaints WSHB - An International Broadcast Station of # letterbox@csms.com The Christian Science Monitor Syndicate, Inc. #technical questions and reports michaelb@wshb.csms.com +1 803 625 4880 # letterbox-tech@csms.com