[Smartmontools-support]Re: smartmontools

[Bruce Allen <ballen@gr...>]

> >Could you please help me to understand this?  In the previous version of
> >the code the amount of data written depended upon what was returned by
> >modesense() in scsiSmartModePage1CHandler.
> 
> No, we copy _into_ the tbuf defined locally in modeselect. Which we
> allocate there and then. And then we send that tBuf to the scsi device as
> a command. MODE_SELECT is sending data to the device, MODE_SENSE is

Of course, you are right.  I was concerned about reading past the end of
pbuf in the memcpy but of course you start at pBuf +  MODE_DATA_HDR_SIZE
and read to pBuf +  MODE_DATA_HDR_SIZE + (CDB_6_MAX_DATA_SIZE -
MODE_DATA_HDR_SIZE)-1, so all is well.

I need more sleep, more coffee, or both...

[By the way, I've apparently introduced an obscure bug in smartd.8.  Any
idea what this means?

smartd.8:245: a node is not allowed in a name
smartd.8:246: a node is not allowed in a name

Cheers,
	Bruce

RE: [Smartmontools-support]smartmontools

[Bruce Allen <ballen@gr...>]

> > > Altough sadly I'm now one step further away from world-wide
> > > recognition. Oh well, such is life.
> > Well there are lots of othere things to fix and improve.  Ask me and I'll
> > provide a list.
> 
> Gimme! :o)

OK.. you asked!

Look at how I do "sleep()" and then interrupt it when I catch SIGUSR1.  
See sleephandler() and search for sleeptime. [Only in the latest CVS
version of smartd.c.]

There must be a better and simpler way of doing this where I just set an
alarm for (say) 30 min from now, and sleep until it goes off, *unless* a
user SIGUSR1 arrives.  I don't like calling sleep every second. It works,
but strikes me as ugly.

Can yoy provide a better solution?

> 
> > (By the way, do you know if we need to fork twice?? I found contradictory
> > opinons when I searched for an answer.)
> 
> Yes, thy shall fork twice. Cannot remember why, though.
> You should do:
> fork
> setsid
> fork
> and then things like chdir("/") umask(0), close(0...2)
> especially the chdir: it's a real pain if you want to unmount the partition
> from which you accidently started the daemon

OK, then the code is ok -- it does fork twice.

Bruce

[Smartmontools-support]Re: smartmontools

[<knan@mo...>]

On Thu, 21 Nov 2002, Bruce Allen wrote:
>Hi Erik,
>> >Also, I was curious if you'd made any further progress in understanding
>> >the scsi segv problem that you were having with both smartmontools and
>> >smartsuite...
>>
>> Well, I just committed a fix... that should answer both your
>> questions, I think :)
>
>Thank you.
>
>There is something that I am confused about in your fix.  The function
>that you fixed (modeselect()) is called in only one place: the function
>scsiSmartModePage1CHandler(), line 457 of scsicmds.c.
>
>When called the last argument of modeselect is a pointer to
>char tBuf[CDB_6_MAX_DATA_SIZE].  Yet with your modification, modeselect
>writes to this array, starting from
>tBuf[CDB_6_HDR_SIZE +MODE_DATA_HDR_SIZE]
>and ending at tBuf[CDB_6_HDR_SIZE+CDB_6_MAX_DATA_SIZE-1].
>So it appears to me that it writes beyond the end of the array char tBuf
>in scsiSmartModePage1CHandler.
>
>Could you please help me to understand this?  In the previous version of
>the code the amount of data written depended upon what was returned by
>modesense() in scsiSmartModePage1CHandler.

No, we copy _into_ the tbuf defined locally in modeselect. Which we
allocate there and then. And then we send that tBuf to the scsi device as
a command. MODE_SELECT is sending data to the device, MODE_SENSE is
reading data from it.

Let's see. We allocate an array of size 256+16 (CDB_6_MAX_DATA_SIZE +
CDB_6_HDR_SIZE) on the stack. Then, at offset 16+12 (CDB_6_HDR_SIZE +
MODE_DATA_HDR_SIZE) in the array, we copy 256-12 (CDB_6_MAX_DATA_SIZE -
MODE_DATA_HDR_SIZE) bytes.

As far as I can tell from simple arithmetic, no overflow of the array. And
the pbuf we read from is CDB_MAX_DATA_SIZE long, 256, allocated in
ModePage1CHandler. So no overflow on reading either, as far as I can tell.

--
Erik I. Bolsø | email: <knan at mo.himolde.no>
The UNIX philosophy basically involves giving you enough rope to
hang yourself.  And then a couple of feet more, just to be sure.

RE: [Smartmontools-support]smartmontools

[Folkert van Heusden <folkert@va...>]

> > Altough sadly I'm now one step further away from world-wide
> > recognition. Oh well, such is life.
> Well there are lots of othere things to fix and improve.  Ask me and I'll
> provide a list.

Gimme! :o)

> (By the way, do you know if we need to fork twice?? I found contradictory
> opinons when I searched for an answer.)

Yes, thy shall fork twice. Cannot remember why, though.
You should do:
fork
setsid
fork
and then things like chdir("/") umask(0), close(0...2)
especially the chdir: it's a real pain if you want to unmount the partition
from which you accidently started the daemon

Re: [Smartmontools-support]Re: UDMA Error Count

[Bruce Allen <ballen@gr...>]

Hi Julian,

> I think the UDMA error can come from having 2 disks on the same cable.
> On my system I had a strenge incompatibility between a Seagate 30G
> (ATA2) and an IBM 30G (DTLA307030) (the IBM is was a bit faulty...)
> that resulted into MANY error reading on both disks. When I removed
> the IBM, everything returned to normal condition. You can clearly see
> that with that incredible 30506 UDMA errors.

I completely agree with this explanation.  Normally theses errors come
from a bad cable, a bad chipset, or a faulty second drive on the same
cable.

Just to be sure, run some drive self-tests if it has self-test capability.

Cheers,
=09Bruce


>    Julian
>=20
> Vendor Specific SMART Attributes with Thresholds:
> ID# ATTRIBUTE_NAME          FLAG     VALUE WORST THRESH TYPE   =20
> WHEN_FAILED RAW_VALUE
>   1 Raw_Read_Error_Rate     0x000e   070   056   025    Old_age      - =
=20
>     2511400
>   3 Spin_Up_Time            0x0002   070   070   000    Old_age      - =
=20
>     0
>   4 Start_Stop_Count        0x0033   100   100   020    Pre-fail     - =
=20
>     151
>   5 Reallocated_Sector_Ct   0x0033   100   100   036    Pre-fail     - =
=20
>     0
>   7 Seek_Error_Rate         0x000f   051   049   030    Pre-fail     - =
=20
>     55777787
>   9 Power_On_Hours          0x0032   095   095   000    Old_age      - =
=20
>     4831
>  10 Spin_Retry_Count        0x0013   100   100   097    Pre-fail     - =
=20
>     0
>  12 Power_Cycle_Count       0x0033   100   100   020    Pre-fail     - =
=20
>     564
> 194 Temperature_Celsius     0x0022   045   071   000    Old_age      - =
=20
>     48
> 195 Hardware_ECC_Recovered  0x001a   085   057   000    Old_age      - =
=20
>     60412153
> 197 Current_Pending_Sector  0x0012   100   100   000    Old_age      - =
=20
>     0
> 198 Offline_Uncorrectable   0x0010   100   100   000    Old_age      - =
=20
>     0
> 199 UDMA_CRC_Error_Count    0x003e   200   001   000    Old_age      - =
=20
>     30506
> 200 Unknown_Attribute       0x0000   099   096   000    Old_age      - =
=20
>     16842752
> 202 Unknown_Attribute       0x0032   100   253   000    Old_age      - =
=20
>     0
>=20
>=20
> __________________________________________________
> Modem offert : 150,92 euros rembours=E9s sur le Pack eXtense de Wanadoo !=
=20
> Haut d=E9bit =E0 partir de 30 euros/mois : http://www.ifrance.com/_reloc/=
w
>=20
>=20
>=20
> -------------------------------------------------------
> This sf.net email is sponsored by:ThinkGeek
> Welcome to geek heaven.
> http://thinkgeek.com/sf
> _______________________________________________
> Smartmontools-support mailing list
> Smartmontools-support@...
> https://lists.sourceforge.net/lists/listinfo/smartmontools-support
>=20

Re: [Smartmontools-support]Strenge outpout

[Bruce Allen <ballen@gr...>]

Hi Julian,

> On one of my hdd, a Seagate 20G (ATA 1) I'm having these very strenge
> results.

I'm confused.  Are you talking about the output of smartctl from below?
It's an ATA-4 disk, and I don't see what's strange.

 The drive is working fine and a self test says OK. The others
> HDD report errors as mentionned by the user having a raid array, maybe
> because I use them with a LVM (Logical Volume Manager)?

I'm sorry, I don't understand your question.

Bruce

> =3D=3D=3D START OF INFORMATION SECTION =3D=3D=3D
> Device Model:     ST320430A                              =20
> Serial Number:    7BX078CE           =20
> Firmware Version: 3.11   =20
> ATA Version is:   4
> ATA Standard is:  Unrecognized. Minor revision code: 0x00
> SMART support is: Available - device has SMART capability.
> SMART support is: Enabled
>=20
> =3D=3D=3D START OF READ SMART DATA SECTION =3D=3D=3D
> SMART overall-health self-assessment test result: PASSED
>=20
> General SMART Values:
> Off-line data collection status: (0x02) Offline data collection activity
>=20
>                                         completed without error.
> Self-test execution status:      (   0) The previous self-test routine
> completed
>                                         without error or no self-test
> has ever=20
>                                         been run.
> Total time to complete off-line=20
> data collection:                 (2222) seconds.
> Offline data collection
> capabilities:                    (0x1b) SMART execute Offline immediate.
>                                         Automatic timer ON/OFF support.
>                                         Suspend Offline collection upon
> new
>                                         command.
>                                         Offline surface scan supported.
>                                         Self-test supported.
> SMART capabilities:            (0x0003) Saves SMART data before entering
>                                         power-saving mode.
>                                         Supports SMART auto save timer.
> Error logging capability:        (0x01) Error logging supported.
> Short self-test routine=20
> recommended polling time:        (   2) minutes.
> Extended self-test routine=20
> recommended polling time:        (  27) minutes.
>=20
> SMART Attributes Data Structure revision number: 10
> Vendor Specific SMART Attributes with Thresholds:
> ID# ATTRIBUTE_NAME          FLAG     VALUE WORST THRESH TYPE   =20
> WHEN_FAILED RAW_VALUE
>   1 Raw_Read_Error_Rate     0x0008   072   068   000    Old_age      - =
=20
>     69733223
>   3 Spin_Up_Time            0x0006   070   070   000    Old_age      - =
=20
>     0
>   4 Start_Stop_Count        0x0013   100   100   020    Pre-fail     - =
=20
>     962
>   5 Reallocated_Sector_Ct   0x0013   100   100   036    Pre-fail     - =
=20
>     0
>   7 Seek_Error_Rate         0x000b   067   060   030    Pre-fail     - =
=20
>     10958911
>   9 Power_On_Hours          0x0012   030   001   000    Old_age      - =
=20
>     7346
>  10 Spin_Retry_Count        0x0013   100   100   097    Pre-fail     - =
=20
>     0
>  12 Power_Cycle_Count       0x0013   099   099   020    Pre-fail     - =
=20
>     1985
> 196 Reallocated_Event_Count 0x0012   100   100   000    Old_age      - =
=20
>     0
> 197 Current_Pending_Sector  0x0030   162   162   000    Old_age      - =
=20
>     18446744073709551615
> 198 Offline_Uncorrectable   0x0030   162   162   000    Old_age      - =
=20
>     18446744073709551615
> 199 UDMA_CRC_Error_Count    0x000a   200   200   000    Old_age      - =
=20
>     1
>=20
> SMART Error Log Version: 1
> No Errors Logged
>=20
> SMART Self-test log, version number 1
> Num  Test_Description    Status                  Remaining=20
> LifeTime(hours)  LBA_of_first_error
> # 1  Extended off-line   Completed                     00%      7344   =
=20
>    =20
> # 2  Short off-line      Completed                     00%      7343   =
=20
>    =20
> # 3  Extended off-line   Interrupted (host reset)      90%       739   =
=20
>    =20
> # 4  Short off-line      Interrupted (host reset)      90%       739   =
=20
>    =20
> # 5  Extended off-line   Interrupted (host reset)      90%       739   =
=20
>    =20
> # 6  Short off-line      Interrupted (host reset)      90%       739   =
=20
>    =20
> # 7  Extended off-line   Interrupted (host reset)      90%       739   =
=20
>    =20
> # 8  Short off-line      Interrupted (host reset)      90%       739   =
=20
>    =20
> # 9  Short off-line      Completed                     00%       697   =
=20
>    =20
> #10  Unknown test        Interrupted (host reset)      150%         0  =
=20
>     =20
> #11  Unknown test        Aborted by host               150%         0  =
=20
>     =20
>=20
>=20
> __________________________________________________
> Modem offert : 150,92 euros rembours=E9s sur le Pack eXtense de Wanadoo !=
=20
> Haut d=E9bit =E0 partir de 30 euros/mois : http://www.ifrance.com/_reloc/=
w
>=20
>=20
>=20
> -------------------------------------------------------
> This sf.net email is sponsored by:ThinkGeek
> Welcome to geek heaven.
> http://thinkgeek.com/sf
> _______________________________________________
> Smartmontools-support mailing list
> Smartmontools-support@...
> https://lists.sourceforge.net/lists/listinfo/smartmontools-support
>=20

RE: [Smartmontools-support]smartmontools

[Bruce Allen <ballen@gr...>]

> > I'd advise keeping our current approach.
> BA> Any comments about this?  I was (eventually) planning to do what your
> BA> patch does (modulo closing ALL fds).  But since daemon() isn't universal
> BA> perhas Erik's point is a good one.
> 
> Yes, I totally agree; keep the original approach.

OK.

> Altough sadly I'm now one step further away from world-wide
> recognition. Oh well, such is life.

Well there are lots of othere things to fix and improve.  Ask me and I'll
provide a list.

(By the way, do you know if we need to fork twice?? I found contradictory 
opinons when I searched for an answer.)

Cheers,
	Bruce

[Smartmontools-support]Re: smartmontools

[Bruce Allen <ballen@gr...>]

Hi Erik,

> >Also, I was curious if you'd made any further progress in understanding
> >the scsi segv problem that you were having with both smartmontools and
> >smartsuite...
> 
> Well, I just committed a fix... that should answer both your
> questions, I think :)

Thank you.

There is something that I am confused about in your fix.  The function
that you fixed (modeselect()) is called in only one place: the function
scsiSmartModePage1CHandler(), line 457 of scsicmds.c.

When called the last argument of modeselect is a pointer to
char tBuf[CDB_6_MAX_DATA_SIZE].  Yet with your modification, modeselect
writes to this array, starting from
tBuf[CDB_6_HDR_SIZE +MODE_DATA_HDR_SIZE]
and ending at tBuf[CDB_6_HDR_SIZE+CDB_6_MAX_DATA_SIZE-1].
So it appears to me that it writes beyond the end of the array char tBuf
in scsiSmartModePage1CHandler.

Could you please help me to understand this?  In the previous version of
the code the amount of data written depended upon what was returned by
modesense() in scsiSmartModePage1CHandler.

Cheers,
	Bruce

[Smartmontools-support]Re: UDMA Error Count

[Julian Vassaux <vassaux@if...>]

Hi
I think the UDMA error can come from having 2 disks on the s=
ame cable.
On my system I had a strenge incompatibility between a Seaga=
te 30G
(ATA2) and an IBM 30G (DTLA307030) (the IBM is was a bit fau=
lty...) that
resulted into MANY error reading on both disks. When I remov=
ed the IBM,
everything returned to normal condition. You can clearly see=
 that with
that incredible 30506 UDMA errors.
   Julian

Vendor Specific SMART Attributes with Thresholds:
ID# ATTRIBUTE_NAME          FLAG     VALUE WORST THRESH TYPE=
   =20
WHEN_FAILED RAW_VALUE
  1 Raw_Read_Error_Rate     0x000e   070   056   025    Old_=
age      - =20
    2511400
  3 Spin_Up_Time            0x0002   070   070   000    Old_=
age      - =20
    0
  4 Start_Stop_Count        0x0033   100   100   020    Pre-=
fail     - =20
    151
  5 Reallocated_Sector_Ct   0x0033   100   100   036    Pre-=
fail     - =20
    0
  7 Seek_Error_Rate         0x000f   051   049   030    Pre-=
fail     - =20
    55777787
  9 Power_On_Hours          0x0032   095   095   000    Old_=
age      - =20
    4831
 10 Spin_Retry_Count        0x0013   100   100   097    Pre-=
fail     - =20
    0
 12 Power_Cycle_Count       0x0033   100   100   020    Pre-=
fail     - =20
    564
194 Temperature_Celsius     0x0022   045   071   000    Old_=
age      - =20
    48
195 Hardware_ECC_Recovered  0x001a   085   057   000    Old_=
age      - =20
    60412153
197 Current_Pending_Sector  0x0012   100   100   000    Old_=
age      - =20
    0
198 Offline_Uncorrectable   0x0010   100   100   000    Old_=
age      - =20
    0
199 UDMA_CRC_Error_Count    0x003e   200   001   000    Old_=
age      - =20
    30506
200 Unknown_Attribute       0x0000   099   096   000    Old_=
age      - =20
    16842752
202 Unknown_Attribute       0x0032   100   253   000    Old_=
age      - =20
    0


__________________________________________________
Modem offert : 150,92 euros rembours=E9s sur le Pack eXtense=
 de Wanadoo !=20
Haut d=E9bit =E0 partir de 30 euros/mois : http://www.ifranc=
e.com/_reloc/w

[Smartmontools-support]Strenge outpout

[Julian Vassaux <vassaux@if...>]

Hi!
On one of my hdd, a Seagate 20G (ATA 1) I'm having these ver=
y strenge
results. The drive is working fine and a self test says OK.
The others HDD report errors as mentionned by the user havin=
g a raid
array, maybe because I use them with a LVM (Logical Volume M=
anager)?
   Julian

=3D=3D=3D START OF INFORMATION SECTION =3D=3D=3D
Device Model:     ST320430A                              =20
Serial Number:    7BX078CE           =20
Firmware Version: 3.11   =20
ATA Version is:   4
ATA Standard is:  Unrecognized. Minor revision code: 0x00
SMART support is: Available - device has SMART capability.
SMART support is: Enabled

=3D=3D=3D START OF READ SMART DATA SECTION =3D=3D=3D
SMART overall-health self-assessment test result: PASSED

General SMART Values:
Off-line data collection status: (0x02) Offline data collect=
ion activity

                                        completed without er=
ror.
Self-test execution status:      (   0) The previous self-te=
st routine
completed
                                        without error or no =
self-test
has ever=20
                                        been run.
Total time to complete off-line=20
data collection:                 (2222) seconds.
Offline data collection
capabilities:                    (0x1b) SMART execute Offlin=
e immediate.
                                        Automatic timer ON/O=
FF support.
                                        Suspend Offline coll=
ection upon
new
                                        command.
                                        Offline surface scan=
 supported.
                                        Self-test supported.
SMART capabilities:            (0x0003) Saves SMART data bef=
ore entering
                                        power-saving mode.
                                        Supports SMART auto =
save timer.
Error logging capability:        (0x01) Error logging suppor=
ted.
Short self-test routine=20
recommended polling time:        (   2) minutes.
Extended self-test routine=20
recommended polling time:        (  27) minutes.

SMART Attributes Data Structure revision number: 10
Vendor Specific SMART Attributes with Thresholds:
ID# ATTRIBUTE_NAME          FLAG     VALUE WORST THRESH TYPE=
   =20
WHEN_FAILED RAW_VALUE
  1 Raw_Read_Error_Rate     0x0008   072   068   000    Old_=
age      - =20
    69733223
  3 Spin_Up_Time            0x0006   070   070   000    Old_=
age      - =20
    0
  4 Start_Stop_Count        0x0013   100   100   020    Pre-=
fail     - =20
    962
  5 Reallocated_Sector_Ct   0x0013   100   100   036    Pre-=
fail     - =20
    0
  7 Seek_Error_Rate         0x000b   067   060   030    Pre-=
fail     - =20
    10958911
  9 Power_On_Hours          0x0012   030   001   000    Old_=
age      - =20
    7346
 10 Spin_Retry_Count        0x0013   100   100   097    Pre-=
fail     - =20
    0
 12 Power_Cycle_Count       0x0013   099   099   020    Pre-=
fail     - =20
    1985
196 Reallocated_Event_Count 0x0012   100   100   000    Old_=
age      - =20
    0
197 Current_Pending_Sector  0x0030   162   162   000    Old_=
age      - =20
    18446744073709551615
198 Offline_Uncorrectable   0x0030   162   162   000    Old_=
age      - =20
    18446744073709551615
199 UDMA_CRC_Error_Count    0x000a   200   200   000    Old_=
age      - =20
    1

SMART Error Log Version: 1
No Errors Logged

SMART Self-test log, version number 1
Num  Test_Description    Status                  Remaining=20
LifeTime(hours)  LBA_of_first_error
# 1  Extended off-line   Completed                     00%  =
    7344   =20
   =20
# 2  Short off-line      Completed                     00%  =
    7343   =20
   =20
# 3  Extended off-line   Interrupted (host reset)      90%  =
     739   =20
   =20
# 4  Short off-line      Interrupted (host reset)      90%  =
     739   =20
   =20
# 5  Extended off-line   Interrupted (host reset)      90%  =
     739   =20
   =20
# 6  Short off-line      Interrupted (host reset)      90%  =
     739   =20
   =20
# 7  Extended off-line   Interrupted (host reset)      90%  =
     739   =20
   =20
# 8  Short off-line      Interrupted (host reset)      90%  =
     739   =20
   =20
# 9  Short off-line      Completed                     00%  =
     697   =20
   =20
#10  Unknown test        Interrupted (host reset)      150% =
        0  =20
    =20
#11  Unknown test        Aborted by host               150% =
        0  =20
    =20


__________________________________________________
Modem offert : 150,92 euros rembours=E9s sur le Pack eXtense=
 de Wanadoo !=20
Haut d=E9bit =E0 partir de 30 euros/mois : http://www.ifranc=
e.com/_reloc/w

RE: [Smartmontools-support]smartmontools

[Folkert van Heusden <folkert@va...>]

> >OK, so the patch as it stands needs a bit of modification.  But it's
still
> >cleaner...
>
> Note: This will be glibc 2.0+ only... libc5 doesn't have the daemon()
> call, according to a quick strings libc.so | grep daemon
>
> And as smartmontools could be interesting for the embedded people w/libc4,
> libc5, uclibc and such, perhaps best not to use too obscure functions?
>
> I'd advise keeping our current approach.
BA> Any comments about this?  I was (eventually) planning to do what your
BA> patch does (modulo closing ALL fds).  But since daemon() isn't universal
BA> perhas Erik's point is a good one.

Yes, I totally agree; keep the original approach.


Folkert


Altough sadly I'm now one step further away from world-wide recognition. Oh
well, such is life.

[Smartmontools-support]Re: smartmontools

[<knan@mo...>]

On Tue, 19 Nov 2002, Bruce Allen wrote:
>It looks like the sourceforge CVS stuff may finally be getting sorted
>out.... though I am still not sure if you have CVS access.  Have you tried
>recently?
>
>Also, I was curious if you'd made any further progress in understanding
>the scsi segv problem that you were having with both smartmontools and
>smartsuite...

Well, I just committed a fix... that should answer both your questions, I
think :)

(megaraid-specific problem - it returns a blank answer to the mode_sense
command - just zeros out the buffer. And a response length of zero is
rather invalid. Worked around it.)

In any event, by reading the megaraid driver, I've discovered that smart
is entirely useless on this device's logical drives, AFAICT. It just
returns "SUCCESS" to all mode_sense commands, no matter what.

Oh well.

--
Erik I. Bolsø | email: <knan at mo.himolde.no>
The UNIX philosophy basically involves giving you enough rope to
hang yourself.  And then a couple of feet more, just to be sure.

[Smartmontools-support]RE: Support for SuSE Linux (fwd)

[Drew J. Como <dcomo@ba...>]

Exactly...

=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D
Drew J. Como             Phone: 631-434-6600
Systems Administrator      Fax: 631-434-7800
dcomo@...         Web: http://www.bascom.com
    Bascom Global Internet Services, Inc.
--------------------------------------------
          "When quality is the goal,
           winning is guaranteed."


-----Original Message-----
From: Stanislav Brabec [mailto:sbrabec@...]
Sent: Thursday, November 21, 2002 10:22 AM
To: Drew J. Como
Cc: 'Fr=E9d=E9ric L. W. Meunier';
smartmontools-support@...
Subject: Re: Support for SuSE Linux (fwd)


On Thu Nov 21 09:54:10 2002, Drew J. Como wrote:
> All,
>
> I am running SuSE 7.x, which has some minor differences
> from the newer 8.x versions.  In this respect, I had to make a slight
> change to the spec file.  (The package _does_ build with the below
> fields in the spec file, however when the RPM is installed, an
> error is thrown.)
>
> I had to comment out the following fields:
>
> #PreReq:       %insserv_prereq %fillup_prereq
>
> % post
> # %{fillup_and_insserv -n smartd}
>
> Removing the above keeps the package functional as well as eliminates
> post install errors.
>

Then you need to add smartd to startup sequences manually. These lines
are for registering init scripts to init engine.

--
Best Regards / S pozdravem,

Stanislav Brabec
pracovn=EDk v=FDvoje softwaru / software developer
---------------------------------------------------------------------
SuSE CR, s. r. o.                             e-mail: sbrabec@...
Drahobejlova 27                               tel: +420 296 542 382
190 00 Praha 9                                fax: +420 296 542 374
Czech Republic                                http://www.suse.cz/

[Smartmontools-support]Re: Support for SuSE Linux (fwd)

[Stanislav Brabec <sbrabec@su...>]

On Thu Nov 21 09:54:10 2002, Drew J. Como wrote:
> All,
> 
> I am running SuSE 7.x, which has some minor differences
> from the newer 8.x versions.  In this respect, I had to make a slight
> change to the spec file.  (The package _does_ build with the below
> fields in the spec file, however when the RPM is installed, an
> error is thrown.)
> 
> I had to comment out the following fields:
> 
> #PreReq:       %insserv_prereq %fillup_prereq
> 
> % post
> # %{fillup_and_insserv -n smartd}
> 
> Removing the above keeps the package functional as well as eliminates
> post install errors.
> 

Then you need to add smartd to startup sequences manually. These lines
are for registering init scripts to init engine.

-- 
Best Regards / S pozdravem,

Stanislav Brabec
pracovník vývoje softwaru / software developer
---------------------------------------------------------------------
SuSE CR, s. r. o.                             e-mail: sbrabec@...
Drahobejlova 27                               tel: +420 296 542 382
190 00 Praha 9                                fax: +420 296 542 374
Czech Republic                                http://www.suse.cz/

[Smartmontools-support]RE: Support for SuSE Linux (fwd)

[Drew J. Como <dcomo@ba...>]

All,

I am running SuSE 7.x, which has some minor differences
from the newer 8.x versions.  In this respect, I had to make a slight
change to the spec file.  (The package _does_ build with the below
fields in the spec file, however when the RPM is installed, an
error is thrown.)

I had to comment out the following fields:

#PreReq:       %insserv_prereq %fillup_prereq

% post
# %{fillup_and_insserv -n smartd}

Removing the above keeps the package functional as well as eliminates
post install errors.

Enjoy :-)

=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D
Drew J. Como             Phone: 631-434-6600
Systems Administrator      Fax: 631-434-7800
dcomo@...         Web: http://www.bascom.com
    Bascom Global Internet Services, Inc.
--------------------------------------------
          "When quality is the goal,
           winning is guaranteed."




-----Original Message-----
From: Stanislav Brabec [mailto:sbrabec@...]
Sent: Thursday, November 21, 2002 4:50 AM
To: Fr=E9d=E9ric L. W. Meunier
Cc: smartmontools-support@...; Drew J. Como
Subject: Re: Support for SuSE Linux (fwd)


On Wed Nov 20 16:08:05 2002, Fr=E9d=E9ric L. W. Meunier wrote:
> I don't know as I don't use SuSE. I'm forwarding it to the
> mailing-list and Stanislav, the developer who works for them.
>
> I don't even know if SuSE uses a "functions" script. You could
> try searching in /etc
>
> ---------- Forwarded message ----------
> Fred,
>
> I was able to build the RPM's for SuSE, however when I
> try and run the daemon, the following messages appear:
>
> /etc/init.d/smartd: /etc/init.d/functions: No such file or directory
> Starting smartd: /etc/init.d/smartd: daemon: command not found
> touch: creating `/var/lock/subsys/smartd': No such file or directory
>
There are SuSE RPMs, binary one is compiled for SuSE 8.1. Smartmontools
will appear oficially in SuSE Linux 8.2.

I have permission to publish SuSE port in the list and Smartmontools
website, so you can put also there.

--
Best Regards / S pozdravem,

Stanislav Brabec
pracovn=EDk v=FDvoje softwaru / software developer
---------------------------------------------------------------------
SuSE CR, s. r. o.                             e-mail: sbrabec@...
Drahobejlova 27                               tel: +420 296 542 382
190 00 Praha 9                                fax: +420 296 542 374
Czech Republic                                http://www.suse.cz/

Re: [Smartmontools-support]Smartd and Email

[Bruce Allen <ballen@gr...>]

> I one of you knowing in which file can I configure my mail server (I
> don't want to use linuxconf, please just point me the file). If needed,

I forgot to say, the daemon uses "mail" to send the mail.  Normally this
in /bin/mail.  If mail is in the PATH environment of the script, shell, or
init file from which smartd is started, and resolves to /bin/mail or any
mail program that understands the "-s" (subject) option and a
comma-delimited address list on the command line, and reads the mail
message from stdin, it should work fine.

Let me know if you have problems.

Cheers,
	Brice

Re: [Smartmontools-support]Smartd and Email

[Bruce Allen <ballen@gr...>]

Hi Julian,

Curious that you should ask this.  About five minutes ago I checked a new
version into CVS that has an additional smartd directive (-m) for the
configuration file. It takes values -3, -2, ... , 2, 3.

Please see documentation.

Anyway, if you set it to a negative number (-1, -2 or -3) it will send a
single test email on startup, as well as emails if problems are found.

With -1 or 1, only a single email warning will be sent if a problem is
found.

With -2 or 2, one email per day will be sent as long as the problem
persists.

With -3 or 3 you get emails at longer and longer intervals: once per day,
then two days, then four days, 8 days, etc.

I haven't done enought testing to generate a release, but you can check it
out of CVS and try it.

Je suis tres curieux de savior si ca marche bien...

Bruce

On Thu, 21 Nov 2002, Julian Vassaux wrote:

> Hi all!
> I'm new to this smart monitoring tool.
> I've seen in the deamon config file that it can send an email when a
> problem occured. Is it possible to test this, to be shure it will work
> fine when needed?
> I one of you knowing in which file can I configure my mail server (I
> don't want to use linuxconf, please just point me the file). If needed,
> I'm running Mandrake 8.
> Thanks !
>    Julian
>=20
>=20
> __________________________________________________
> Modem offert : 150,92 euros rembours=E9s sur le Pack eXtense de Wanadoo !=
=20
> Haut d=E9bit =E0 partir de 30 euros/mois : http://www.ifrance.com/_reloc/=
w
>=20
>=20
>=20
> -------------------------------------------------------
> This sf.net email is sponsored by:ThinkGeek
> Welcome to geek heaven.
> http://thinkgeek.com/sf
> _______________________________________________
> Smartmontools-support mailing list
> Smartmontools-support@...
> https://lists.sourceforge.net/lists/listinfo/smartmontools-support
>=20

[Smartmontools-support]Smartd and Email

[Julian Vassaux <vassaux@if...>]

Hi all!
I'm new to this smart monitoring tool.
I've seen in the deamon config file that it can send an emai=
l when a
problem occured. Is it possible to test this, to be shure it=
 will work
fine when needed?
I one of you knowing in which file can I configure my mail s=
erver (I
don't want to use linuxconf, please just point me the file).=
 If needed,
I'm running Mandrake 8.
Thanks !
   Julian


__________________________________________________
Modem offert : 150,92 euros rembours=E9s sur le Pack eXtense=
 de Wanadoo !=20
Haut d=E9bit =E0 partir de 30 euros/mois : http://www.ifranc=
e.com/_reloc/w

[Smartmontools-support]Smartmontools Support for SuSE Linux

[Bruce Allen <ballen@gr...>]

The SuSE 8.1 RPMs provided by Stanislav Brabec have been posted in the
smartmontools file download area.  This will simplify installation for
SuSE users.

Re: [Smartmontools-support]smartmontools

[Bruce Allen <ballen@gr...>]

Folkert,

> >OK, so the patch as it stands needs a bit of modification.  But it's still
> >cleaner...
> 
> Note: This will be glibc 2.0+ only... libc5 doesn't have the daemon()
> call, according to a quick strings libc.so | grep daemon
> 
> And as smartmontools could be interesting for the embedded people w/libc4,
> libc5, uclibc and such, perhaps best not to use too obscure functions?
> 
> I'd advise keeping our current approach.

Any comments about this?  I was (eventually) planning to do what your
patch does (modulo closing ALL fds).  But since daemon() isn't universal
perhas Erik's point is a good one.

Cheers,
	Bruce

Re: [Smartmontools-support]smartmontools and Darwin/MacOS X

[Bruce Allen <ballen@gr...>]

Hi Peter,

> Just a little feedback. I downloaded the 5.0-45 sources last night 
> just for fun & hacked them into building on MacOS X 
> (10.2.2/Kernel6.2/xnu-344.12.2).

Cool!

> Anyways - it compiles, builds and runs after a series of nasty hacks! 
> However .... it appears that the virtual raw disk drivers that the 
> kernel produces (/dev/rdiskx) supports only an extremely narrow set of 
> ioctls. I had a good dig through other sources (including Darwin) to 
> see if I could discover anything but didn't have any luck.

As far as I know, and as you've figured out, it's just the commands in
atacmds.c and scsicmds.c that send ioctls which need to be ported.  
Everything else is ANSI/Posix/GNU C.  And there are really only a few
commands needed.

> I know that if you approach Apple (disclaimer: I develop diagnostics @
> Apple)  they'll tell you that you should be using the IOKit which
> would entail an almost complete re-write, so I'm a bit bummed ........

Hmmm, I don't know enough to ask intelligent questions.... but isn't it
just the commands in atacmds.c and scsicmds.c that would need a complete
rewrite?  Doesn't sound *so* bad...

I'm looking for a developer for a Darwin/MacOS "port" (ideally, just
integrated into the exising code base).  See 

http://sourceforge.net/people/?group_id=64297
and
http://sourceforge.net/people/viewjob.php?group_id=64297&job_id=11410

Are you interested?

Thanks for the feedback!

Bruce

Re: [Smartmontools-support]smartmontools

[<knan@mo...>]

On Thu, 21 Nov 2002, Erik Inge Bolsø wrote:
>On Thu, 21 Nov 2002, Bruce Allen wrote:
>>OK, so the patch as it stands needs a bit of modification.  But it's still
>>cleaner...
>
>Note: This will be glibc 2.0+ only... libc5 doesn't have the daemon()
>call, according to a quick strings libc.so | grep daemon
>
>And as smartmontools could be interesting for the embedded people w/libc4,
>libc5, uclibc and such, perhaps best not to use too obscure functions?

Example: I can guess that smartmontools would be interesting for Axis'
linux people for their DVR. That box runs some homebrew embedded linux
system from flash memory, and stores months worth of surveillance pictures
on 2-4 ide disks. Monitoring failures of those disks with smartmon
sounds worthwhile. I've worked with a few other Axis boxes, which use
Linux on the Etrax processor, kernel 2.0 or 2.4, and certainly no glibc
(in perhaps 8 MB flash? no way)

So I'd advise keeping our current approach.

--
Erik I. Bolsø | email: <knan at mo.himolde.no>
The UNIX philosophy basically involves giving you enough rope to
hang yourself.  And then a couple of feet more, just to be sure.

Re: [Smartmontools-support]smartmontools

[<knan@mo...>]

On Thu, 21 Nov 2002, Bruce Allen wrote:
>> >I do have one question, though.  (This question has been in the TODO file
>> >for some time now!)
>> >
>> >The existing daemon_init() closes ALL open file descriptors, not just
>> >those associated with stdin, stdout, and stderr.  This is important
>> >because a logic bug in smartd might leave open fds from disks, and syslog
>> >might also leave open fds.
>> >
>> >Do I still need to do this closing of "other" fds by hand?  Or will daemon
>> >close all open fds, not just 0,1,2?
>>
>> According to the FreeBSD manpage for daemon, it will close and reopen
>> stdin, stdout and stderr as /dev/null.
>>
>> http://www.gsp.com/cgi-bin/man.cgi?section=3&topic=daemon
>>
>> So it will certainly not close any dangling fds we might be leaving
>> around.
>
>OK, so the patch as it stands needs a bit of modification.  But it's still
>cleaner...

Note: This will be glibc 2.0+ only... libc5 doesn't have the daemon()
call, according to a quick strings libc.so | grep daemon

And as smartmontools could be interesting for the embedded people w/libc4,
libc5, uclibc and such, perhaps best not to use too obscure functions?

I'd advise keeping our current approach.

--
Erik I. Bolsø | email: <knan at mo.himolde.no>
The UNIX philosophy basically involves giving you enough rope to
hang yourself.  And then a couple of feet more, just to be sure.

[Smartmontools-support]smartmontools and Darwin/MacOS X

[Peter Cassidy <pcassidy@ma...>]

Hi there,

	Just a little feedback. I downloaded the 5.0-45 sources last night 
just for fun & hacked them into building on MacOS X 
(10.2.2/Kernel6.2/xnu-344.12.2).

	Anyways - it compiles, builds and runs after a series of nasty hacks! 
However .... it appears that the virtual raw disk drivers that the 
kernel produces (/dev/rdiskx) supports only an extremely narrow set of 
ioctls. I had a good dig through other sources (including Darwin) to 
see if I could discover anything but didn't have any luck. I know that 
if you approach Apple (disclaimer: I develop diagnostics @ Apple)  
they'll tell you that you should be using the IOKit which would entail 
an almost complete re-write, so I'm a bit bummed ........

	Anyway - hope this is of some use. If I get a bit more time, I'll 
re-visit it.

Regards,

Pete Cassidy

Re: [Smartmontools-support]smartmontools

[Bruce Allen <ballen@gr...>]

> >I do have one question, though.  (This question has been in the TODO file
> >for some time now!)
> >
> >The existing daemon_init() closes ALL open file descriptors, not just
> >those associated with stdin, stdout, and stderr.  This is important
> >because a logic bug in smartd might leave open fds from disks, and syslog
> >might also leave open fds.
> >
> >Do I still need to do this closing of "other" fds by hand?  Or will daemon
> >close all open fds, not just 0,1,2?
> 
> According to the FreeBSD manpage for daemon, it will close and reopen
> stdin, stdout and stderr as /dev/null.
> 
> http://www.gsp.com/cgi-bin/man.cgi?section=3&topic=daemon
> 
> So it will certainly not close any dangling fds we might be leaving
> around.

OK, so the patch as it stands needs a bit of modification.  But it's still
cleaner...

Bruce