[Orca-users] Re: Re: Segmentation Fault with 1.35

[alegrand at wallace.com]

> Kyle,
> 	I am attaching a fixed version of 1.35.  You need to use it to get
> devices stats from A1000 and A3500 raid arrays. 
>             Would you please confirm that  c0t6d0  is a CDROM and that the
> next disk on the controller is c0t11d0.  Since,
> 	that is the bug that I just fixed.
> 
> Alan
> 
>  <<orcallator.se-1.35b.txt>> 
> From: "J. Kyle Cummings III" <jkcummi at y...
> </group/orca-users/post?protectID=0452022191820381920362181641402440630780
> 48234051197>> 
> Date: Fri Sep 13, 2002 2:46 pm
> Subject: Re: [orca-users] Re: Segmentation Fault with 1.35
> 
> I discovered when I was testing this problem that one of the six Solaris 7
> servers was corrupted. I then got this from one that worked properly.
> 
> I was able to comment out the USE_RAWDISK statement and got it to work.
> 
> I have attached the disk information from the system which I got the trace
> from.
> 
> Thanks!
> 
> -Kyle
> 
> --- adlegrand <alegrand at w...
> </group/orca-users/post?protectID=2431281921081270311722232031021471302481
> 44234189209171188199>> wrote:
> > 
> > 
> > YuuKk,
> > 
> > This system is having a problem with the RAW disk code
> > 
> > The raw disk code is seeing disks sd{6,11,12,28,29,49 150-162}
> > 
> > The SE disk routines does not see sd150-162. 
> > 
> > 
> > There is code in th disk_map section that uses the first disk
> > on the system to Identify when it has reached the disk slices
> > part of the int the GLOBAL_DISK_INFO array. 
> > 
> > sd6 is the first disk mentioned in the DISK_INFO array but it
> > does not have any slices (probably because it is a cdrom and
> > is not mounted.)
> > 
> > Kyle,
> > 
> > Would you please indentify all of the disks mentioned
> > about so that I can figure out how to stop the code from addressing
> > unallocated memory. ( Probably, I am going to add a second test
> > "second_name" to pick up the first two disks.
> > 
> > 
> > Alan
> > 
> > 
> > 
> > 
> > 
> > 
> > 
> > 
> > 
> > 
> > --- In orca-users at y...
> </group/orca-users/post?protectID=1652122192370131620900381092191141640710
> 48139>, "J. Kyle Cummings III" <jkcummi at y...
> </group/orca-users/post?protectID=045202219182038192036218164036129208>>
> wrote:
> > > Ok, I think I am narrowing the scope.
> > > 
> > > After I re-installed orca-0.27b2, I copied the following files into
> > > /usr/local/lib:
> > > 
> > > orcallator.se-1.35.txt
> > > orcallator.cfg-1.35.txt
> > > 
> > > I then backed up the old se and cfg and file renamed the 1.35 ones 
> > to the
> > > .se and .cfg files.
> > > 
> > > When I start orcallator now, it seg faults.
> > > 
> > > I have attached a text file with the new trace. It looks a bit 
> > different
> > > now...
> > > 
> > > -Kyle
> > > 
> > > --- jkcummi <jkcummi at y...
> </group/orca-users/post?protectID=045202219182038192036218164036129208>>
> wrote:
> > > > I think I have narrowed the problem down to just ONE of my 
> > Solaris 7 
> > > > systems. I am going to go over it with a fine tooth comb today. 
> > It 
> > > > is possible that I am missing a patch or something.
> > > > 
> > > > I went back and re-installed orca-0.27b2 on the 6 systems which 
> > were 
> > > > having problems. 5 of the six started working.
> > > > 
> > > > I am going to put in the 1.35 updates next to see if that makes 
> > them 
> > > > break.
> > > > 
> > > > -Kyle
> > > > 
> > > > --- In orca-users at y...
> </group/orca-users/post?protectID=1652122192370131620900381092191141640710
> 48139>, "jkcummi" <jkcummi at y...
> </group/orca-users/post?protectID=045202219182038192036218164036129208>>
> wrote:
> > > > > Ok, I did some new testing today. I tried to use SE 3.0 but it 
> > > > does 
> > > > > not support Solaris 7. I also tried to use SE 3.3 but it does 
> > not 
> > > > > support 32 bit. (We run Solaris 7 32 bit for application 
> > support 
> > > > > reasons)
> > > > > 
> > > > > I guess I am stuck with using SE 3.2.1.
> > > > > 
> > > > > Is there any additional information I can collect which might 
> > point 
> > > > > the SE folks at the root cause?
> > > > > 
> > > > > Thanks for the help in this...
> > > > > 
> > > > > -Kyle
> > > > > 
> > > > > --- In orca-users at y...
> </group/orca-users/post?protectID=1652122192370131620900381092191141640710
> 48139>, Blair Zajac <blair at o...
> </group/orca-users/post?protectID=014128113009127135112082190036>> wrote:
> > > > > > adlegrand wrote:
> > > > > > > 
> > > > > > > Kyle,
> > > > > > > 
> > > > > > > You seem to have uncovered a bug in the SE 
> > toolkit. 
> > > > Not 
> > > > > the
> > > > > > > raw disk code. :( You have no idea how glad I am about 
> > that. 
> > > > It
> > > > > > > does however strongly resemble the errors that sometimes 
> > > > occured 
> > > > > in
> > > > > > > earlier versions of the RAW DISK code.
> > > > > > > 
> > > > > > > These occured when the code was trying to identify a raw 
> > device 
> > > > > like
> > > > > > > sd27 which did not exist in the structure GLOBAL_disk_info
> > > > > > > [].short_name. It ended up walking off the end of the 
> > array 
> > > > into
> > > > > > > unallocated memory.
> > > > > > > 
> > > > > > > You might try it with an earlier version of the SE toolkit. 
> > To 
> > > > see
> > > > > > > if this solve your problem.
> > > > > > 
> > > > > > Alan,
> > > > > > 
> > > > > > That's great news!
> > > > > > 
> > > > > > If we could figure out where this error occurs, we could send 
> > in 
> > > > a 
> > > > > bug
> > > > > > report to the SE developers.
> > > > > > 
> > > > > > Kyle,
> > > > > > 
> > > > > > You may want to try SE 3.3.
> > > > > > 
> > > > > > Best,
> > > > > > Blair
> > > > > > 
> > > > > > -- 
> > > > > > Blair Zajac <blair at o...
> </group/orca-users/post?protectID=014128113009127135112082190036>>
> > > > > > Web and OS performance plots - <http://www.orcaware.com/orca/>
> > > > 
> > > > 
> > > 
> > > 
> > > 
-------------- next part --------------
An embedded and charset-unspecified text was scrubbed...
Name: orcallator.se-1.35b.txt
URL: </pipermail/orca-users/attachments/20020913/f9e0128c/attachment.txt>