gertjan@txsil.UUCP (01/13/85)
/* Written 10:15 am Jan 3, 1985 by bradley!brad in txsil:net.unix-wizards */
/* ---------- "11/73,2.9,large amounts of output,he" ---------- */
Configuration:
PDP 11/73, 1MEG, Emulex cs01/16 lines, 2 RL02's (RLV controler)
2.9BSD software
Problem:
when running 'yes jfdljfldjflsjfdsjflsdlfjdsjfsljdsdjlfjsljfds'
or any program that outputs large amounts of data to the terminal
and running several terminals ( > 2) together, the programs start
spitting out info that looks like the kernel stuff, i.e. garbage chars.
If I touch the control-S (^S), and stop the output, then control-Q,
the output starts up ok again. Note, that the problem can come back
and that it doesn't happen to all terminals. If I start with 10
running, by the next morning only 2 or 3 are left ok.
If have UCB_CLIST defined and undefined.
If you have any thoughts at all as to where to look let me know.
thanks inadvance.
Bradley Smith UUCP: {cepu,ihnp4,noao,uiucdcs}!bradley!brad
Text Processing ARPA: cepu!bradley!brad@UCLA-LOCUS
Bradley University PH: (309) 676-7611 Ext. 446
Peoria, IL 61625
/* End of text from txsil:net.unix-wizards */kwthomas@uokmet.UUCP (01/14/85)
>Configuration: > PDP 11/73, 1MEG, Emulex cs01/16 lines, 2 RL02's (RLV controler) > 2.9BSD software >Problem: > when running 'yes jfdljfldjflsjfdsjflsdlfjdsjfsljdsdjlfjsljfds' > or any program that outputs large amounts of data to the terminal > and running several terminals ( > 2) together, the programs start > spitting out info that looks like the kernel stuff, i.e. garbage chars. > > If I touch the control-S (^S), and stop the output, then control-Q, > the output starts up ok again. Note, that the problem can come back > and that it doesn't happen to all terminals. If I start with 10 > running, by the next morning only 2 or 3 are left ok. > > If have UCB_CLIST defined and undefined. We have strange problems for a while with ^S and ^Q. When 3 or more people did a ^S to stop output, our system (PDP 11/44 with 2.9BSD) would freeze dead in its tracks. Checking the PC indicated that the operating system was spinning around in the idle() loop! Do a ^Q at any of the "offending" terminals would cause the system to come alive with no effects, other than loss of time on the clock (verified by checking clock time before performing tests). As it turned out, somewhere along the way, I had turned off "UCB_CLIST". When I defined it, the ^S problem went away. Kevin W. Thomas Univ. of Oklahoma School of Meteorology Norman, OK 73019 UUCP: ...!ctvax!uokvax!uokmet!kwthomas