FW: [Orca-dev] Re: One stray system with disk problems

[Jon Tankersley]

-----Original Message-----
From: Jon Tankersley [mailto:jon.tankersley at eds.com] 
Sent: Saturday, July 03, 2004 11:33 PM
To: 'Blair Zajac'; 'Paul Haldane'; 'Dmitry Berezin'
Subject: RE: [Orca-dev] Re: One stray system with disk problems


I think I found it...
It is a bit nasty and I don't know why it happens, but...
In orcallator.se 1.36 and in (1.37) there is a section of code...  It
doesn't check to see if GLOBAL_disk_info[j].short_name is nil or not.  It
is, so the strcpy dies.
      for (j=0; j<MAX_RAWDISKS; ++j) {
        strcpy(short_name, GLOBAL_disk_info[j].short_name);

I just put an if in place before the strcpy:
        if (GLOBAL_disk_info[j].short_name == nil) { break; }

And it seems to progress.  At least I finally got an output file. Dmitry,
this the kind of thing you've seen? 


So, for further diagnosis:
--- orcallator.se-1.36.plus_tape        2002-12-10 15:11:06.869996000 -0600
+++ orcallator.se-1.36.txt.tmp          2004-07-03 23:13:20.499724000 -0500
@@ -637,6 +637,7 @@
     if (strncmp(RAW_disk[i].short_name, "st", 2) != 0) {
       if (strncmp(RAW_disk[i].short_name, "fd", 2) != 0) {
         for (j=0; j<MAX_RAWDISKS; ++j) {
+          if (GLOBAL_disk_info[j].short_name == nil) { break; }
           strcpy(short_name, GLOBAL_disk_info[j].short_name);
           if (j > 0) {
             if (strncmp(first_name, short_name, first_len) == 0) {

--- orcallator.se-1.37.txt      2003-07-11 14:43:02.000000000 -0500
+++ orcallator.se.1.37.tmp      2004-07-03 23:10:38.779723000 -0500
@@ -692,6 +692,7 @@
     if (strncmp(RAW_disk[i].short_name, "st", 2) != 0 &&
         strncmp(RAW_disk[i].short_name, "fd", 2) != 0) {
       for (j=0; j<MAX_RAWDISKS; ++j) {
+        if (GLOBAL_disk_info[j].short_name == nil) { break; }
         strcpy(short_name, GLOBAL_disk_info[j].short_name);
         if (j > 0) {
           short_name_length = strlen(short_name);

-----Original Message-----
From: orca-dev-bounces+jon.tankersley=eds.com at orcaware.com
[mailto:orca-dev-bounces+jon.tankersley=eds.com at orcaware.com] On Behalf Of
Blair Zajac
Sent: Saturday, July 03, 2004 7:38 PM
To: Paul Haldane
Cc: orca-dev at orcaware.com
Subject: Re: [Orca-dev] Re: One stray system with disk problems


Paul Haldane wrote:

>>From: Jon Tankersley <jon.tankersley at eds.com>
>>Subject: [Orca-dev] One stray system with disk problems....
>>
>>Regardless of RAWDISK setting or MAXCOLUMNS/MAXDISK, we dump core due 
>>to segfault....
>>
> 
> ...
> 
>>strcpy(short_name<ssd2,h>, GLOBAL_disk_info[35].short_name<(nil)>)^M
>>Segmentation Fault^M
> 
> 
> This _may_ be related to the problem that I've encountered recently
> with a couple of our systems (Solaris 9, DiskSuite, SAN disks and 
> multipath drivers).  Needed to increase the size of long_name in the 
> RawDisk structure.
> 
> Diffs of orcallator.se at end.
> 
> Paul

Hi Paul,

Thanks for the diff.

Turns out that fix was already applied to orcallator.se 1.39:

+// Version 1.39:   Feb 18, 2003 Burkhardt Schopp
+//                              B.Schopp at gmx.de
+//                              Modified struct RawDisk long_name object to
+//                              48 characters to handle T3 disks names.

However, that change only increased it to 48 characters.  Does it need to be
64 
characters long?

Regards,
Blair

-- 
Blair Zajac <blair at orcaware.com>
Plots of your system's performance - http://www.orcaware.com/orca/
_______________________________________________
Orca-dev mailing list
Orca-dev at orcaware.com http://www.orcaware.com/mailman/listinfo/orca-dev