[ale] suddenly finding computer 'seized-up'....ATTN:DowHurst-2

Thu Mar 23 10:06:42 EST 2006

Dow,
Thank you for your continued attention  :-)

Please see interspersed reply below:

----- Original Message ----- 
From: "Dow_Hurst" <dhurst at mindspring.com>
To: ale at ale.org
To: "Atlanta Linux Enthusiasts" <ale at ale.org>
Sent: Wednesday, March 22, 2006 10:11 PM
Subject: Re: [ale] suddenly finding computer 'seized-up'....ATTN:DowHurst

> Googling parts of the error message indicated that errors on hard drives
show drives starting to fail as the major reason.  Only a few pages were
found so this isn't a commonly posted error message.  Or, my Google search
was lacking.  Anyway, your showing two symptoms of possible a possible
hardware problem starting to manifest.  I'm thinking if you check your boxes
interior and carefully blow it free of dust and check the fan intakes and
outputs that could relieve some heat stress.  Otherwise, I'm thinking you
will experience more erratic errors until the problem gets bad enough to
really diagnose.

I took the sides off the box when all this started up, so, I guess it's not
interior temp anyway.

The box is a few years old, runs about 900 mhz, has about 320MB ram and has
a 200GB drive which is not run from the onboard sockets but as a pseudo-scsi
connection from a Promise card. When I tried to switch it to the motherboard
sockets I got even more errors so I'm assuming it's the HD, a Seagate that
has about 5 more yrs. warranty. So, I guess I'll send it back. Fortunately,
I've got a CD backup [untested] so I hope I'll be able to recover the bulk
of installations but have since installed much more  :-(  and doubt I'll be
able to reciver that.

I left the box on last night to see what would happen...and it was seized
this AM, so fsck'd and only /var had a small discrepancy. Why does nothing
respond after a seizure if only the HD is at fault, i.e. why doesn't the
mouse, screen or kbd respond b/c the HD is 'gone' ? I tried to ping the box
but 'host unavailable'. I haven't gotten ssh working from another box, yet.
I get the applicable error...'protocol mismatch' with telnet and I'm feelin'
my way in installing ssh.

>
> The modem settings under FreeBSD are like in Linux where the
initialization string that configures the modem is sent by a chat script.
It sets up the modem and dials the number, after authentication it passes
control to pppd.  wvdial is a simple utility that simplifies the whole dial
out process and is very popular to use.  I was looking at this link:
>
> http://www.freebsd.org/doc/en/books/handbook/dialout.html
>
> and thinking the /etc/remote file has your modem init settings stored
there.
>
> I still wonder if you have varying voltage problems from either old
hardware, high internal temps, or bad line voltage.  As voltage drops
current rises and stuff starts acting erratic.  Are you on a UPS?  Have you
had any brownouts lately from the power company?

I'm on a ECM Powercom UPS. I had a power interruption for a few hours last
week [squirrel blew up a transformer just outside the house] in which I
failed to connect with the box before the UPS lost it's battery. I don't
know what bad voltage the UPS is capable of handling. I had an APC but went
with the less expensive unit that has less battery shutdown time.

This AM I installed a new HD controller,...one of those PCI Maxtor
pseudo-scsi cards for IDE drives but still the same HD errors about writing,
etc...on bootup. I guess it's the HD. Sure hate to go through all the grief
of sending it back to Seagate, not to mention a couple of months of
installations to be redone, if it's not the HD.

Appreciatively,

Courtney
Hope these suggestions/thoughts help out,
> Dow
>
>
> -----Original Message-----
> >From: Courtney Thomas <cc.thomas at earthlink.net>
> >Sent: Mar 21, 2006 9:45 AM
> >To: Atlanta Linux Enthusiasts <ale at ale.org>
> >Subject: Re: [ale] suddenly finding computer 'seized-up'....ATTN:DowHurst
> >
> >Dow,
> >
> >Thank you so much for your attention with this problem.
> >
> >I'll practice the described remote access, hopefully, before the problem
> >recurs. This is a skill I obviously should've obtained long ago.
> >
> >Re: seizure
> >
> >1-I've used "xset s on" for years for my 'screensaver' which just
> >darkens the screen, so I don't think that's "it".
> >
> >2-I'm not aware that I'm using either RTS/CTS or XON/XOFF, but I guess
> >somewhere it's configured, but I don't know where. I dialup/logon to the
> >web with a script which has no mention of this. It's a Hayes 56K modem
> >whose speed is set to 115200 in the script. But I've also used this
> >script for years without system seizure, so conclude it's probably not
> >that either.
> >
> >3-After a cold boot this morn, I DID get the following error multiple
> >times with different LBA #s:
> >
> >ad4:FAILURE - WRITE_DMA status=51 <READY,DSC,ERROR> error=4<ABORTED>
> >dma=0x06 LBA=44xxxxx
> >
> >I got a lot of LBA=44xxxxx, the xxxxx s being different numbers.
> >
> >After not getting an error-free boot, I fsck'd the ad4 partitions,
> >rebooted and all seemed OK.
> >
> >4-Regarding the FreeBSD recommendation of increasing the cp4ticks from 4
> >to 40, I assume that after this is done, recompilation of what,... the
> >entire kernel,... is necessary to implement the change ?
> >
> >5-and finally,... how can I tweak my modem settings, which I assume
> >pertains to the rts/cts xon/xoff configuration, if that's the problem ?
> >
> >Most appreciatively,
> >
> >Courtney
> >
> >
> >Dow Hurst wrote:
> >> Courtney,
> >> I've just assumed everyone on the list would be proficient in the use
of
> >> ssh.  Maybe we should have a tutorial on this most necessary of
tools?!!
> >>
> >> SSH is the secure form of what telnet provides.  Most all Linux distros
> >> come with this preinstalled in the default selections and will run the
> >> sshd server on boot.  Normally port 22 is where the sshd server lives.
> >> The ssh client will use hostnames or IPs to connect.  So if you have
> >> another Linux machine on your network you can:
> >>
> >> ssh hosedmachine
> >>
> >> and you should get a password prompt back.  If not then:
> >>
> >> telnet hosedmachine 22
> >>
> >> will get you a string of text if the sshd server is running.  Type some
> >> gibberish and press Enter to be disconnected from the server.  It
should
> >> tell you what server version your machine is running.  There are lot's
> >> of uses for ssh for sysadmins.  O'Reilly has an updated book on SSH.
If
> >> your system is Sys/V based then /etc/init.d/sshd start should start the
> >> secure shell server up.  There are cools things like key based
> >> passwordless logins that are highly secure or rsync backup scripts thru
> >> ssh.  It's a vital tool for sysadmins to encrypt and protect the login
> >> and data passed between machines that ftp and telnet expose to the
world.
> >>
> >> My thoughts on your problem are:
> >> 1. that your screensaver is kicking with powermanagement but has some
> >> problem or bug and won't "wake up"
> >>
> >> 2. A hardware interrupt problem can lock up your kernel so no one
really
> >> is at home.  A little googling revealed this thread as a start:
> >> http://lists.freebsd.org/pipermail/freebsd-bugs/2003-May/000685.html
> >> http://lists.freebsd.org/pipermail/freebsd-bugs/2003-May/000687.html
> >>
> >> Maybe your modem settings need tweaking instead of what they recommend?
> >> Are you using hardware flow control or software flow control?
> >> Dow
> >>
> >>
> >>
> >> Courtney Thomas wrote:
> >>
> >>>Steven,
> >>>
> >>>Thank you for your suggestions.
> >>>
> >>>I guess telnet would be the easiest to try, huh ?
> >>>
> >>>I've no experience with the other two and both machines are on my local
> >>>LAN so ssh is not required I assume, but any further elaboration on how
> >>>to use these 3 tools would be gratefully accepted.
> >>>
> >>>Cordially,
> >>>Courtney
> >>>
> >>>
> >>>
> >>>Steven A. DuChene wrote:
> >>>
> >>>
> >>>>If the box is still running, have you tried accessing it via
ssh/rsh/telnet
> >>>
> >>>>from anoter system prior to just pulling the plug for a hard reset?
> >>>
> >>>>-----Original Message-----
> >>>>
> >>>>
> >>>>
> >>>>>From: Courtney Thomas <cc.thomas at earthlink.net>
> >>>>>Sent: Mar 20, 2006 2:26 PM
> >>>>>To: Atlanta Linux Enthusiasts <ale at ale.org>
> >>>>>Subject: [ale] suddenly finding computer 'seized-up'....
> >>>>>
> >>>>>Recently several times I've left my computer downloading for extended
> >>>>>periods to return to a frozen machine, i.e. not working are the
mouse,
> >>>>>keyboard or monitor but the box itself is still running; and, of
course,
> >>>>>the only way to resume is...a hard reset. Fortunately, this has not
> >>>>>resulted in disaster. I've fsck'd all the partitions before rebooting
> >>>>>and, seemingly nothing critical has been lost.
> >>>>>
> >>>>>I should mention that often but not always, when I try to print from
> >>>>>Mozilla 1.7.7 [not Firebird, but rather Mozilla/5 Gecko 20051115] I
get
> >>>>>a core dump. This usually occurs after Mozilla has been running
awhile,
> >>>>>not when first started. I wonder if all this is from a Mozilla memory
> >>>>>leak and if so, how can I tell ? Or should I just upgrade Mozilla
[HOW
> >>>>>?] to remove this suspect, or is the recent version equally likely to
> >>>>>repeat.
> >>>>>
> >>>>>I also ran memtest86 and discovered nothing.
> >>>>>
> >>>>>I'm intermittently getting a stream of sio [sio1, irq=3] overflows,
the
> >>>>>mouse is on com1 and modem on com2.
> >>>>>
> >>>>>Nevertheless, clearly, I need to resolve this ASAP.
> >>>>>
> >>>>>A suitable algorithm [or a pointer to same] pertaining to
hard&soft-ware
> >>>>>resolution...would be appreciated.
> >>>>>
> >>>>>Gratefully,
> >>>>>
> >>>>>Courtney
> >>>>>
> >>>>
> >>>>_______________________________________________
> >>>>Ale mailing list
> >>>>Ale at ale.org
> >>>>http://www.ale.org/mailman/listinfo/ale
> >>>>
> >>>>
> >>>
> >>>_______________________________________________
> >>>Ale mailing list
> >>>Ale at ale.org
> >>>http://www.ale.org/mailman/listinfo/ale
> >>>
> >>>
> >>
> >>
> >> _______________________________________________
> >> Ale mailing list
> >> Ale at ale.org
> >> http://www.ale.org/mailman/listinfo/ale
> >>
> >_______________________________________________
> >Ale mailing list
> >Ale at ale.org
> >http://www.ale.org/mailman/listinfo/ale
>
>
> No sig.
> _______________________________________________
> Ale mailing list
> Ale at ale.org
> http://www.ale.org/mailman/listinfo/ale