[Orca-users] Seg Fault on 2 Sunfire 280's

[Big Chuck]

Jaouich.Cyril at hydro.qc.ca wrote:
> Hi Big Chuck,
> 
>         I had the same problem with a Solaris 8, I ended up commenting 
> the parameter #define USE_RAWDISK in the orcallator.se
> 
>         It has worked ever since.
> 
>         Thanks
> 
> Cyril Jaouich (Consultant Unix)
> ------------------------------------------
> Support technique des infrastructures Unix
> ------------------------------------------
> Tél: 514-840-3000x5527
> 
> 
> 
> -----Message d'origine-----
> De : Big Chuck [mailto:ccarson at echeeba.com]
> Envoyé : mercredi 30 juillet, 2003 21:40
> À : Blair Zajac
> Cc : Orca Mailing List
> Objet : Re: [Orca-users] Seg Fault on 2 Sunfire 280's
> 
> 
> Blair Zajac wrote:
>  > Big Chuck wrote:
>  >
>  >>Blair Zajac wrote:
>  >>
>  >>>Big Chuck wrote:
>  >>
>  >><snip>
>  >>
>  >>>>I am at a loss. I can find no different on these machines other 
> than the
>  >>>>fact that the working machines on running Veritas Foundation Suite and
>  >>>>the two databases are running Database Edition for Oracle. (thus not a
>  >>>>whole hell of a lot of difference)
>  >>>>
>  >>>>Any ideas?
>  >>>
>  >>>
>  >>>Look for any oddly named filesystem devices in /dev and /devices.  They
>  >>>may need to be removed.
>  >>>
>  >>>Also, you can run orcallator.se with -d and you'll get debugging output
>  >>>which will be useful to see where it crashes.
>  >>>
>  >>>Best,
>  >>>Blair
>  >>>
>  >>
>  >>Odd, I just installed it onto an E4500 that also has Database Edition
>  >>for Oracle with the /dev/odm and /dev/qlog devices entry it works on
>  >>that box.
>  >>
>  >>Hmmm, I am at a loss. Unfortunately the main system I wanted to monitor
>  >>was one of the 280's that it does not work on. =(
>  >
>  >
>  > Those two devices won't cause the crash.
>  >
>  > Did you run the find?
>  >
>  > Did you run se -d?
>  >
>  > Best,
>  > Blair
>  >
> 
> I ran the find and it did not find any files with spaces. When I run se
> -d orcallator.se, I get several pages of output but here are the last
> several lines:
> 
> strcpy(short_name<ssd1,c>, GLOBAL_disk_info[15].short_name<ssd1,d>)
> if (j<15> > <0>)
> if (strncmp(first_name<sd6>, short_name<ssd1,d>, first_len<3>) == <0>)
> if (strcmp(RAW_disk[2].short_name[2]<9>, short_name<ssd1,d>) == <0>)
> j++;
> strcpy(short_name<ssd1,d>, GLOBAL_disk_info[16].short_name<ssd1,e>)
> if (j<16> > <0>)
> if (strncmp(first_name<sd6>, short_name<ssd1,e>, first_len<3>) == <0>)
> if (strcmp(RAW_disk[2].short_name[2]<9>, short_name<ssd1,e>) == <0>)
> j++;
> strcpy(short_name<ssd1,e>, GLOBAL_disk_info[17].short_name<ssd1,g>)
> if (j<17> > <0>)
> if (strncmp(first_name<sd6>, short_name<ssd1,g>, first_len<3>) == <0>)
> if (strcmp(RAW_disk[2].short_name[2]<9>, short_name<ssd1,g>) == <0>)
> j++;
> strcpy(short_name<ssd1,g>, GLOBAL_disk_info[18].short_name<(nil)>)
> Segmentation Fault
> 
> -Chuck
> 
> 
> 
> 
> _______________________________________________
> Orca-users mailing list
> Orca-users at orcaware.com
> http://www.orcaware.com/mailman/listinfo/orca-users
> 
> 
> ------------------------------------------------------------------------
> 
> _______________________________________________
> Orca-users mailing list
> Orca-users at orcaware.com
> http://www.orcaware.com/mailman/listinfo/orca-users



You had the problem on all Solaris 8 systems? Mine is isolated to two 
280R's running Solaris 8 64-bit and Veritas Database Edition for Oracle. 
  However, at the same time, it works on an E4500 running Datasbe 
Edition for Oracle.

I just tried what you suggested and it indeeed works now. What impact 
does disabling this setting have? Also, did you comment that entry out 
or change it to 0?

I wonder if we can compare notes and determine what the root cause is?

Thanks,
CC