headers
Bruce Allen | 1 Jan 2003 08:59
Picon
Favicon

Re: Hard lockup for smartctl and Promise 20265(+ide-taskfile?)

Mark, Georg,

Every few days, I do a google search to see what's turned up that's
relevant to smartmontools.  Here's something of relevance to our
promise-related thread.  Mark, I think it suggests an answer to your
question (that yes, the command below WILL fail) but I'd still like to see
what happens when Georg has a chance to try it...

http://groups.google.com/groups?q=smartctl+-freshmeat+-slackware&hl=en&lr=&ie=UTF-8&oe=UTF-8&scoring=d&selm=200212201338.45492.marvin%40synapse.net&rnum=5

Cheers,
	Bruce

On Mon, 30 Dec 2002, Mark Lord wrote:

> Mmm.. if this is true, then the following should ALSO fail:
> 
>      cat /proc/ide/hd*/identify
> 
> Does it?
> 
> Cheers
> -- 
> Mark Lord
> Real-Time Remedies Inc.
> mlord <at> pobox.com
> 
> 
> 
>
(Continue reading)

Permalink | Reply | 6 comments |
headers
Bruce Allen | 1 Jan 2003 09:04
Picon
Favicon

Re: Hard lockup for smartctl and Promise 20265(+ide-taskfile?)

Another thread of relevance:

http://groups.google.com/groups?hl=en&lr=&ie=UTF-8&oe=UTF-8&threadm=fa.mu2pj6v.5624bf%40ifi.uio.no&rnum=7&prev=/groups%3Fq%3Dsmartctl%2B-freshmeat%2B-slackware%26hl%3Den%26lr%3D%26ie%3DUTF-8%26oe%3DUTF-8%26scoring%3Dd%26selm%3Dfa.mu2pj6v.5624bf%2540ifi.uio.no%26rnum%3D7

Cheers,
	Bruce

> Mark, Georg,
> 
> Every few days, I do a google search to see what's turned up that's
> relevant to smartmontools.  Here's something of relevance to our
> promise-related thread.  Mark, I think it suggests an answer to your
> question (that yes, the command below WILL fail) but I'd still like to see
> what happens when Georg has a chance to try it...
> 
> http://groups.google.com/groups?q=smartctl+-freshmeat+-slackware&hl=en&lr=&ie=UTF-8&oe=UTF-8&scoring=d&selm=200212201338.45492.marvin%40synapse.net&rnum=5
> 
> Cheers,
> 	Bruce
> 
> On Mon, 30 Dec 2002, Mark Lord wrote:
> 
> > Mmm.. if this is true, then the following should ALSO fail:
> > 
> >      cat /proc/ide/hd*/identify
> > 
> > Does it?
> > 
> > Cheers
> > --
(Continue reading)

Permalink | Reply | 0 comments |
headers
Georg Acher | 2 Jan 2003 21:04
Picon
Favicon

Re: Hard lockup for smartctl and Promise 20265(+ide-taskfile?)

On Wed, Jan 01, 2003 at 01:59:39AM -0600, Bruce Allen wrote:
> Mark, Georg,
> 
> Every few days, I do a google search to see what's turned up that's
> relevant to smartmontools.  Here's something of relevance to our
> promise-related thread.  Mark, I think it suggests an answer to your
> question (that yes, the command below WILL fail) but I'd still like to see
> what happens when Georg has a chance to try it...
> 
> http://groups.google.com/groups?q=smartctl+-freshmeat+-slackware&hl=en&lr=&ie=UTF-8&oe=UTF-8&scoring=d&selm=200212201338.45492.marvin%40synapse.net&rnum=5

Sounds very similar to my setup. I have also disabled the Promise-Bios,
since it takes so long to detect the disks. When I'm back home, I'll try
whether enabling it will help or not.

--

-- 
         Georg Acher, acher <at> in.tum.de         
         http://wwwbode.in.tum.de/~acher
         "Oh no, not again !" The bowl of petunias          

-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Welcome to geek heaven.
http://thinkgeek.com/sf
Permalink | Reply | 4 comments |
headers
Georg Acher | 6 Jan 2003 15:43
Picon
Favicon

Re: Hard lockup for smartctl and Promise 20265(+ide-taskfile?)

On Thu, Jan 02, 2003 at 09:04:51PM +0100, Georg Acher wrote:
> On Wed, Jan 01, 2003 at 01:59:39AM -0600, Bruce Allen wrote:
> > Mark, Georg,
> > 
> > Every few days, I do a google search to see what's turned up that's
> > relevant to smartmontools.  Here's something of relevance to our
> > promise-related thread.  Mark, I think it suggests an answer to your
> > question (that yes, the command below WILL fail) but I'd still like to see
> > what happens when Georg has a chance to try it...
> > 
> > http://groups.google.com/groups?q=smartctl+-freshmeat+-slackware&hl=en&lr=&ie=UTF-8&oe=UTF-8&scoring=d&selm=200212201338.45492.marvin%40synapse.net&rnum=5
> 
> Sounds very similar to my setup. I have also disabled the Promise-Bios,
> since it takes so long to detect the disks. When I'm back home, I'll try
> whether enabling it will help or not.

Ok, back at home: /cat/proc/hde/identify also crashes the machine. But more
important: Enabling the Promise-Bios (aka ATA100-Bios) again solves all
troubles. So it's not taskfile related in any way, it's a incomplete
initialization of the Promise chip in the kernel driver.

Thanks for all the support, now I know that my drives are still healthy ;-)

--

-- 
         Georg Acher, acher <at> in.tum.de         
         http://wwwbode.in.tum.de/~acher
         "Oh no, not again !" The bowl of petunias          

-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
(Continue reading)

Permalink | Reply | 3 comments |
headers
Bruce Allen | 6 Jan 2003 16:56
Picon
Favicon

Re: Hard lockup for smartctl and Promise 20265(+ide-taskfile?)

Georg,

I'm glad that this is sorted out.

I want to modify the smartmontools WARNING file to help users overcome
this problem.  I'll put in "Enable the Promise-BIOS (ATA100-BIOS) or the
machine will hang". But please tell me if this is not the right thing to
say, or if I should say more.

Cheers,
	Bruce

On Mon, 6 Jan 2003, Georg Acher wrote:

> On Thu, Jan 02, 2003 at 09:04:51PM +0100, Georg Acher wrote:
> > On Wed, Jan 01, 2003 at 01:59:39AM -0600, Bruce Allen wrote:
> > > Mark, Georg,
> > > 
> > > Every few days, I do a google search to see what's turned up that's
> > > relevant to smartmontools.  Here's something of relevance to our
> > > promise-related thread.  Mark, I think it suggests an answer to your
> > > question (that yes, the command below WILL fail) but I'd still like to see
> > > what happens when Georg has a chance to try it...
> > > 
> > > http://groups.google.com/groups?q=smartctl+-freshmeat+-slackware&hl=en&lr=&ie=UTF-8&oe=UTF-8&scoring=d&selm=200212201338.45492.marvin%40synapse.net&rnum=5
> > 
> > Sounds very similar to my setup. I have also disabled the Promise-Bios,
> > since it takes so long to detect the disks. When I'm back home, I'll try
> > whether enabling it will help or not.
>
(Continue reading)

Permalink | Reply | 2 comments |
headers
Georg Acher | 6 Jan 2003 17:56
Picon
Favicon

Re: Hard lockup for smartctl and Promise 20265(+ide-taskfile?)

On Mon, Jan 06, 2003 at 09:56:28AM -0600, Bruce Allen wrote:
> Georg,
> 
> I'm glad that this is sorted out.

Me too :-)

> I want to modify the smartmontools WARNING file to help users overcome
> this problem.  I'll put in "Enable the Promise-BIOS (ATA100-BIOS) or the
> machine will hang". But please tell me if this is not the right thing to
> say, or if I should say more.

That should be enough. Alernatively (the luxury version ;-), you can check
/for proc/ide/pdc202xx in the program, issue a warning and continue after a
small delay to give X the chance for a screen sync...

--

-- 
         Georg Acher, acher <at> in.tum.de         
         http://wwwbode.in.tum.de/~acher
         "Oh no, not again !" The bowl of petunias          

-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Welcome to geek heaven.
http://thinkgeek.com/sf
Permalink | Reply | 1 comment |
headers
Bruce Allen | 6 Jan 2003 18:17
Picon
Favicon

Re: Hard lockup for smartctl and Promise 20265(+ide-taskfile?)

Georg,

> That should be enough. Alernatively (the luxury version ;-), you can check
> /for proc/ide/pdc202xx in the program, issue a warning and continue after a
> small delay to give X the chance for a screen sync...

Can you send me a diff or code patch?

Bruce

-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Welcome to geek heaven.
http://thinkgeek.com/sf
Permalink | Reply | 0 comments |
headers
Tim Southerwood | 9 Jan 2003 20:47
Picon
Picon
Favicon

Thanks - and some random ideas

Hello All, (Especially the maintainers, whom
this message is mostly targeted at).

First, please let me introduce myself - I'm a sysadmin at
one of the University of London colleges.

Secondly, please read everything that follows as
"just my ideas, maybe I would do it like...".
I am not one of the maintainers of this tool,
so I have no right to criticise or pass
judgement. Everything I will say is intended
to be random ideas or stuff I think would be
useful to me and maybe someone else. Maybe
you hate my ideas or patches - that's fine.

smartmontools is a great package - I'm really impressed
so far having tried it out on 200 odd PCs. Thank
you for producing this :-)

Anyway, formalities over - here's some thoughts
I had:

1) I like smartd and it's email support - I have error
notifications going directly to a web support system
(wreq). However I decided it would be useful to add
more verbosity to the email message. Specifically 
I wanted to append the output of "smartctl -a" to the end
so that I had some context. To this end, I knocked up
a quick and ikky patch called "verboseemail.patch".
(Continue reading)

Permalink | Reply | 2 comments |
headers
Tim Southerwood | 9 Jan 2003 21:20
Picon
Picon
Favicon

Patch to add output of "smartctl -a" to smartd's mail message if -M enabled

Here's an attempt for a new feature. It's partially tested
(works for me).

Note: Patch is against version 5.0.49

Purpose: To add a dump of "smartctl -a <device>" to any mail
message that smartd sends.

Why? I find it useful to see warning messages with full context.
I cannot always log into the machine smartd has flagged up
because user may have booted it to MS Windows.

Caveats: Only uses /bin/mail (mailx package) as was originally used, 
paths to /bin/mail and /usr/sbin/smartctl hardcoded in smartd.h . 
No configuration option to enable - it's always on.

Implementation notes:

I felt I should rewrite the mail sending bit not to use system() to
avoid problems with the output of smartctl -a confusing the
shell and maybe introducing undesireable behaviour.

smartd.h and smartd.c now have helper functions added:

1 - int sendmail (const char* to, const char* subject, const char* body) 
actually does the job, using a fork/exec/pipe
approach shamelessly lifted from the comp.unix.programmers (sp?)
FAQ.

2 - int writeall (const int fd, const char* buf, const ssize_t buflen)
(Continue reading)

Permalink | Reply | 0 comments |
headers
Tim Southerwood | 9 Jan 2003 21:27
Picon
Picon
Favicon

Patch to Makefile to allow building RPM as non root user

Note: Patch against release code 5.0.49

Purpose: To be applied in RPM spec file, to allow the RPM
to be built as non root user (ie don't try chown).

Why? Useful on NFS hosted source trees, where being root doesn't
work well. Also good practice not to have to be root to build software
packages.

It's a bit trivial really - but I had it so I though I may as well send
it here :-)

Best wishes

Tim Southerwood

******************* CUT HERE *********************************
diff -Naur smartmontools-5.0-49.orig/Makefile smartmontools-5.0-49/Makefile
--- smartmontools-5.0-49.orig/Makefile  Fri Dec  6 09:30:51 2002
+++ smartmontools-5.0-49/Makefile       Tue Jan  7 17:04:22 2003
 <at>  <at>  -95,21 +95,21  <at>  <at> 
        /bin/gzip -c smartd.conf.5 > smartd.conf.5.gz
        rm -f $(DESTDIR)/usr/share/man/man8/smartctl.8
        rm -f $(DESTDIR)/usr/share/man/man8/smartd.8
-       install -m 755 -o root -g root -D smartctl         $(DESTDIR)/usr/sbin/smartctl
-       install -m 755 -o root -g root -D smartd           $(DESTDIR)/usr/sbin/smartd
-       install -m 755 -o root -g root -D smartd.initd     $(DESTDIR)/etc/rc.d/init.d/smartd
-       install -m 644 -o root -g root -D smartctl.8.gz    $(DESTDIR)/usr/share/man/man8/smartctl.8.gz
-       install -m 644 -o root -g root -D smartd.8.gz      $(DESTDIR)/usr/share/man/man8/smartd.8.gz
-       install -m 644 -o root -g root -D smartd.conf.5.gz $(DESTDIR)/usr/share/man/man5/smartd.conf.5.gz
(Continue reading)

Permalink | Reply | 0 comments |

Gmane