ccvj@lure.latrobe.edu.au (05/16/91)
It just grows and grows! Has anyone come across this before? Mips M-120, RISC/os 4.52, a student machine (I say this only because I thought it was the students!) I have a problem with talkd in that the processes never disappear. Usually, I just go in and kill them, but today I found another aspect to it. I don't know whether there is something with talk, file protections and whether there is a separate problem with biff or RISCwindows (just recently installed), but this is what happens. -r-xr-xr-x 1 bin bin 163840 Jan 31 12:05 /bsd43/bin/talk crwx------ 1 ccvj cc 15, 4 May 16 11:29 /dev/ttyq4 # who root tty0 May 6 17:21 1:59 29384 (console == tty0) ccvj ttyq4 May 10 13:37 0:12 4694 csixw ttyq1 May 15 11:42 1:38 21326 root ttyq3 May 15 16:40 . 26078 ...so I try kill -9 21326, because this user really is not on the system anymore. Actually, I have an idle process killer which executes approx every 15 minutes which is supposed to be doing this , but the process does not actually exist, so root is mailed that info, I myself (not root) have an xterm session going (ttyq4), and I keep getting advised that new mail has arrived for root. Every time mail arrives (rather I am notified by biff) another talkd process happens. Before being notified of mail for root (ie first ps): 6 processes root 26475 154 0 16:57:33 ? 0:00 talkd root 26341 1 0 16:49:28 ? 0:00 talkd root 26447 1 0 16:55:32 ? 0:00 talkd root 26421 1 0 16:53:31 ? 0:00 talkd root 26476 26475 0 16:57:33 ? 0:00 talkd root 26387 1 0 16:51:30 ? 0:00 talkd root 26523 26092 4 16:59:06 ttyq3 0:00 grep talk After notification - 7 processes root 26525 154 0 16:59:34 ? 0:00 talkd root 26341 1 0 16:49:28 ? 0:00 talkd root 26447 1 0 16:55:32 ? 0:00 talkd root 26421 1 0 16:53:31 ? 0:00 talkd root 26476 1 0 16:57:33 ? 0:00 talkd root 26387 1 0 16:51:30 ? 0:00 talkd root 26526 26525 0 16:59:34 ? 0:00 talkd root 26528 26092 2 16:59:53 ttyq3 0:00 grep talk Yet later later.. - 9 processes root 26765 26092 4 17:15:05 ttyq3 0:00 grep talk root 26609 1 0 17:01:35 ? 0:00 talkd root 26647 1 0 17:05:37 ? 0:00 talkd root 26341 1 0 16:49:28 ? 0:00 talkd root 26447 1 0 16:55:32 ? 0:00 talkd root 26421 1 0 16:53:31 ? 0:00 talkd root 26476 1 0 16:57:33 ? 0:00 talkd root 26387 1 0 16:51:30 ? 0:00 talkd root 26628 1 0 17:03:36 ? 0:00 talkd root 26526 1 0 16:59:34 ? 0:00 talkd Of course, I have now stopped mail being sent by cron for the idle process killer, but I really do not understand what is happening. Could someone please explain the connection between biff and talk (if there is one). There appear to be at least two separate problems: 1. the talkds never dying and 2. why my xterm is notified for root's mail plus 3? the relationship between biff and talk. If the answer is something obvious, please email me, as my ego is fairly fragile at the moment. Thanks in advance! ------ Vicki Jordan La Trobe University ccvj@lure.latrobe.edu.au Bundoora, Vic 3083 Australia.
kts@michael.udev.cdc.com (kathy t skow x-2923) (05/16/91)
In article <1991May16.120940.1@lure.latrobe.edu.au>, ccvj@lure.latrobe.edu.au writes: |> It just grows and grows! |> Has anyone come across this before? Mips M-120, RISC/os 4.52, a |> student machine (I say this only because I thought it was the students!) |> I have a problem with talkd in that the processes never disappear. |> Usually, I just go in and kill them.... We, too, are seeing talkd processes that do not go away. We are running EP/IX 1.3.2 on a variety of CD 4000s, and I have seen this on the CD 4680 (equivalent to the MIPS 6280) and the CD 4360 -- and it is probably occurring on the other platforms, as well. Does anyone have any clues on this one?? Thanks. ------------------------------------------------------------------------ Kathy Skow | email: kts@udev.cdc.com Open Systems Operations | phone: 612-482-2923 Control Data Corporation | fax: 612-482-4203 "If perception is reality, then we'd better figure out how to quantify and track it" ...unknown
lgy@phys.washington.edu (Laurence Yaffe) (05/17/91)
kts@michael.udev.cdc.com (kathy t skow x-2923) writes:
-We, too, are seeing talkd processes that do not go away. We are
-running EP/IX 1.3.2 on a variety of CD 4000s, and I have seen this
-on the CD 4680 (equivalent to the MIPS 6280) and the CD 4360 -- and
-it is probably occurring on the other platforms, as well.
-Does anyone have any clues on this one??
No clues, but this is definitely a problem under RISCos 4.51 and 4.52.
--
--------------------------------------------------------------------------
Laurence G. Yaffe Internet: lgy@newton.phys.washington.edu
University of Washington Bitnet: yaffe@uwaphast.bitnet
josevela@mtecv2.mty.itesm.mx (Jose Angel Vela Avila) (05/19/91)
>We, too, are seeing talkd processes that do not go away. We are >running EP/IX 1.3.2 on a variety of CD 4000s, and I have seen this >on the CD 4680 (equivalent to the MIPS 6280) and the CD 4360 -- and >it is probably occurring on the other platforms, as well. >Does anyone have any clues on this one?? We had this problem too in an RS/6000 with the ugly AIX 3.1 with 0,1,2 and 3 'maintenance' updates (just ugly patches).. We report it and send us a new talkd that seems to work well Jose A. Vela A.