2.1.7 locking up often :(

2.1.7 locking up often :(

Post by Mikhail Teter » Sat, 20 Dec 1997 04:00:00



Hello! I'm trying to figure out what's going on with heavily loaded
underpowered web-server. Starting about 2 month ago, it locks up every
once in a while... Does not respond to the network, does not write to
the logs, hitting ENTER on the console never gives you the next line.
Cold RESET is required... Adding more memory (up to 96Mb) helped it
stay up for one day longer, but froze again. I was running the

        systat 1 -vmstat

on it all the time. Here is what the "corpse's watch is stuck on":

-------------- Screen of "systat 1 -vmstat" ----------------------------------
    2 users    Load  3.49  2.17  2.17                  Thu Dec 18 13:35

Mem:KB    REAL            VIRTUAL                     VN PAGER  SWAP PAGER
        Tot   Share      Tot    Share    Free         in  out     in  out
Act   17164    1692   294708   252356    3320 count    
All   95372   20832  1234180   876276         pages    
                                                       49 cow    Interrupts
Proc:r  p  d  s  w    Csw  Trp  Sys  Int  Sof  Flt    141 zfod   1300 total
     9     1 54       154  397  494 1352   45  374  12824 wire    100 clk0 irq0
                                                    74736 act     128 rtc0 irq8
18.0%Sys  34.5%Intr 46.9%User  0.5%Nice  0.0%Idl     3080 inact       fdc0 irq6
|    |    |    |    |    |    |    |    |    |       3176 cache   993 wdc0 irq14
=========+++++++++++++++++>>>>>>>>>>>>>>>>>>>>>>      144 free     79 ed1 irq5
                                                          daefr
Namei         Name-cache    Proc-cache                 75 prcfr
    Calls     hits    %     hits     %                    react
      323      313   97        1     0                    pdwake
                                                          pdpgs
Discs  wd0  wd1                                           intrn
seeks   24   73                                      6503 buf
xfers   24   73
 blks  162  335
 msps  6.0  8.2
-------------------------------------------------------------------------

Am I right, and it is the wdc0 that's*? 993 interrupts per second...
34.5% of a CPU in the IRQ... Adding memory reduced paging, I guess, that's
why it stayed longer -- my hypothesis.

The CPU is AMD, reported as ``i486DX''. Runs on 100MHz . For whatever
reason, the PC has it's IDE controller on ISA bus, even though it also
has PCI.

Any comments? Thanks!

        -mi

--
        "Windows for dummies"

 
 
 

1. Talk refused to talk:(:(:(:(

Hello there people:)...
  I had linux kernel 1.2.8 installed in my system. I have a SLIP/PPP
connection using either dip (for SLIP) or pppd/chat (for PPP).
  Now, Most of my network stuff are working fine without any problem at all
except talk. Talk works fine for local chatting. No problem at all. The
problem is when ever I call out to other machine (using the SLIP/PPP line),
nothing has ever happen. It got as far as Checking invitation stuff. Never
more.
  Just out of curiosity I ran netstate to check the udp connection whenever
I have talk running. Running locally (talk that is) I got output as follow:

Active Internet connections
Proto Recv-Q Send-Q Local Address          Foreign Address        (State)
User
udp        0      0 lucifer.newpaltz.:1328 *:*                    

Which I presumed opened by talk (only one here because the other side hadn't
responded to the call yet).  I also get the same thing whenever I call
someone outside my system to talk. Of course, it never got to the other
system (I was waiting on the other system using telnet for the talk request,
never got there. And I had made sure that mesg was y).

Now when I call *IN* from other system (telnet to another system and run
talk over there to my system), I have (on my system) output as follow:

Active Internet connections
Proto Recv-Q Send-Q Local Address          Foreign Address        (State)
User

And also from netstat -a, I could find talk/ntalk were still listening.
Seems to me the call from outside didn't get through.

Now, since I haven't done anything to configure talk specifically, is there
anything I have to do to make this connection possible. Since I can't talk
to outside nor can I receive talk request from other hosts, I'm under the
assumption that my talk program somehow is not configured properly.

Any configuration file I have to check?
man, faq, howto, etc to refer to this kind of error?
May be replace my talk/ntalk/talkd/ntalkd with something else and where to
find it/them?
Does it depend on something in the kernel that I have to enable during the
compile time?
Wrong version altogether with the kernel?
Is it possible that talk needs long time to make the connection? (None of my
other stuff that I know of behave like this).
May be my inetd is broken?
If so, how do I check for it?
And where to find the latest version of it?

And while I'm at it, what's the latest version of talk?
Where to find it?
Is there a better chatting program that's backward compatible with talk?
Meaning that the program has more features than talk but at the same time
can be used to communicate with other regular talk program whenever the
program is not available on the other host.

--
Send comment and reply to:



2. bad disk blocks

3. Often freeze ups after motherboard swap

4. Newbie Compiler Question

5. makes failing too often... :(

6. Connecting remote Kerberos site from Solaris 7 machine...

7. Lock ups with SCO and Caps lock

8. Solaris 8 CD-ROM mounted to the root file system

9. my trashed hard drive. :( :( :(

10. Shape Ups,Men's Shape Ups,Men's Skechers Shape Ups - new styles!

11. System Locks Up TOO Often

12. Problem with AGP ViRGE/GX2 board - unexpected lock ups

13. keyboard lock-ups