marc@jahangir.UUCP (Marc Rossner) (06/20/91)
We are running ISC2.2 with 2.2.1 fixes on a Northgate 486 with X11/1.2 and the X server will occasionally die with the error: "xterm: Error 50, errno 11: No more processes" We have NPROC at 200 and MAXUP at 60 so there is no way we are really running out of processes. TFM also says errno 11 could be "a system call failed because of insufficient memory or swap space." We have 8 meg RAM and 16 meg swap -- does this have anything to do with anything? u386mon currently reports we are using 98% of memory and 5% of swap. While there is probably no causal relationship we did start to see this error only after we got TCP/IP 1.2 with NFS running so there are a lot of daemons running but no where near 200 processes or 60 for a non-root user. Marc Rossner jahangir!marc@uunet.uu.net
rfp@uucs1.UUCP (rfp) (06/22/91)
In article <559@jahangir.UUCP> marc@jahangir.UUCP (Marc Rossner) writes: >We are running ISC2.2 with 2.2.1 fixes on a Northgate 486 with X11/1.2 >and the X server will occasionally die with the error: > >"xterm: Error 50, errno 11: No more processes" > We have seen the same problem on a generic 25Mhz 386 system with an Adaptec SCSI controller, Paradise 8514, etc. I can reliably reproduce it by pulling up an xterm from our ISC/386 system, displaying on an X server on another system, an then ignoring it. After a while, it will croak as above. -- Roger F. Powell UUCS inc. Phoenix, Az ncar!noao!asuvax!hrc!uucs1!rfp sun!sunburn!gtx!uucs1!rfp
greg@halexii.uucp (Gregory F. Hogg) (06/24/91)
In article <515@uucs1.UUCP> rfp@uucs1.UUCP (Roger F. Powell) writes: >In article <559@jahangir.UUCP> marc@jahangir.UUCP (Marc Rossner) writes: >>We are running ISC2.2 with 2.2.1 fixes on a Northgate 486 with X11/1.2 >>and the X server will occasionally die with the error: >> >>"xterm: Error 50, errno 11: No more processes" >> >We have seen the same problem on a generic 25Mhz 386 system with an >Adaptec SCSI controller, Paradise 8514, etc. I can reliably reproduce >it by pulling up an xterm from our ISC/386 system, displaying on an >X server on another system, an then ignoring it. After a while, it >will croak as above. > I had this problem and 'fixed' it by setting TMOUT=0. I.e. I told korn shell not to timeout and this problem stoped. -- Greg Hogg COMPANY : HOGG'S SOFTWARE