on 06-02-2009 11:08 AM
Hi Folks,
Its on a new PI 7.1 build
I set the working directory in a saposcol.pfl profile file and started saposcol after downloading latest one from marketplace and copied to kernel
DIR_PERF=/usr/sap/<SID>/ccms
As sidadm :- <kernelpath>/saposcol pf=<profilepath>/saposcol.pfl
saposcol is running fine, All the other OS monitors are getting reported properly ( such as CPU, memory, swap etc)
But i can see only root Filesystem, other local and NFS filesystems are not visible under RZ20 or ST06
Operating System
<hostname>_<SID>_00
Filesystems
/
Name /
Freespace 3646 MB, Green 02.06.2009 , 12:02:52
Percentage_Used 28 % , Green 02.06.2009 , 12:02:52
Regards
Shanmuga
Edited by: Shanmuga on Jun 2, 2009 3:40 PM
Edited by: Shanmuga on Jun 2, 2009 3:42 PM
> I set the working directory in a saposcol.pfl profile file and started saposcol after downloading latest one from marketplace and copied to kernel
>
> DIR_PERF=/usr/sap/<SID>/ccms
> As sidadm :- <kernelpath>/saposcol pf=<profilepath>/saposcol.pfl
>
> saposcol is running fine, All the other OS monitors are getting reported properly ( such as CPU, memory, swap etc)
did you execute "./saproot.sh <SID>" in the kernel directory? What's the output of
ls -l `which saposcol`
as user <sid>adm?
Markus
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
Hi Markus,
$ sudo /sapmnt/???/exe/uc/sunx86_64/saproot.sh ???
Preparing /usr/sap/???/SYS/exe/uc/sunx86_64/brbackup ...
Preparing /usr/sap/???/SYS/exe/uc/sunx86_64/brarchive ...
Preparing /usr/sap/???/SYS/exe/uc/sunx86_64/brconnect ...
Preparing icmbnd ...
icmbnd.new does not exist - skipped
done
ls -l 'saposcol'
-rwxrwxr-x 1 ???adm sapsys 1906304 Mar 10 20:18 saposcol
p.s - SID substituted with ???
Edited by: Shanmuga on Jun 3, 2009 12:14 PM
> $ sudo /sapmnt/???/exe/uc/sunx86_64/saproot.sh ???
> Preparing /usr/sap/???/SYS/exe/uc/sunx86_64/brbackup ...
> Preparing /usr/sap/???/SYS/exe/uc/sunx86_64/brarchive ...
> Preparing /usr/sap/???/SYS/exe/uc/sunx86_64/brconnect ...
>
> Preparing icmbnd ...
>
> icmbnd.new does not exist - skipped
> done
icmbnd is a central part of the kernel, it HAS to exist.
> ls -l 'saposcol'
>
> -rwxrwxr-x 1 ???adm sapsys 1906304 Mar 10 20:18 saposcol
The permissions are wrong, it seems sudo does not do its job in this case.
Usually you have to
cdexe (this is an alias for the kernel directory)
su -
./saproot.sh <SID>
Markus
Hi Markus,
Thanks for your replies
The icm* files does exist
/usr/sap/???/SYS/exe/uc/sunx86_64
hostname:???adm> ls -l icmb*
-rwsx- 1 root sapsys 1265696 May 18 19:34 icmbnd
-rwxr-xr-x 1 ???adm sapsys 1355912 May 19 20:57 icmbnd.new
I have got the saposcol owner and permission changed
hostname:???adm> ls -l sapos*
-rwsr-x--- 1 root sapsys 1906304 Mar 10 20:18 saposcol
and restarted saposcol - SAME ISSUE, only root filesystem details are coming up
Edited by: Shanmuga on Jun 5, 2009 1:24 PM
> I have got the saposcol owner and permission changed
>
>
> hostname:???adm> ls -l sapos*
> -rwsr-x--- 1 root sapsys 1906304 Mar 10 20:18 saposcol
That looks good now.
Really strange....
What do you get get when you execute (as <sid>adm)
saposcol -o > output.txt
and then search for "Disk". Do you see the disks there?
Markus
From the output.txt
Under Disk Data
I have many number of lines - like the ones below ( only few lines i have copied here)
sd1 util: 0% queue: 0 wait: 0 serv: 0 kbsec: 0 opsec: 0
sd2 util: 0% queue: 0 wait: 0 serv: 0 kbsec: 0 opsec: 0
md11 util: 0% queue: -2147483648 wait: 0 serv: 0 kbsec: 0 opsec: -2147483648
md12 util: 0% queue: -2147483648 wait: 0 serv: 0 kbsec: 0 opsec: -2147483648
md10 util: 0% queue: -2147483648 wait: 0 serv: 0 kbsec: 0 opsec: -2147483648
sd46 util: 0% queue: -2147483648 wait: 0 serv: 0 kbsec: 0 opsec: -2147483648
sd47 util: 0% queue: -2147483648 wait: 0 serv: 0 kbsec: 0 opsec: -2147483648
sd48 util: 0% queue: -2147483648 wait: 0 serv: 0 kbsec: 0 opsec: -2147483648
sd49 util: 0% queue: -2147483648 wait: 0 serv: 0 kbsec: 0 opsec: -2147483648
sd50 util: 0% queue: -2147483648 wait: 0 serv: 0 kbsec: 0 opsec: -2147483648
sd51 util: 0% queue: -2147483648 wait: 0 serv: 0 kbsec: 0 opsec: -2147483648
sd52 util: 0% queue: -2147483648 wait: 0 serv: 0 kbsec: 0 opsec: -2147483648
sd53 util: 0% queue: -2147483648 wait: 0 serv: 0 kbsec: 0 opsec: -2147483648
sd54 util: 0% queue: 0 wait: 0 serv: 0 kbsec: 5 opsec: 2
sd55 util: 0% queue: -2147483648 wait: 0 serv: 0 kbsec: 0 opsec: -2147483648
sd56 util: 0% queue: -2147483648 wait: 0 serv: 0 kbsec: 0 opsec: -2147483648
sd43 util: 0% queue: -2147483648 wait: 0 serv: 0 kbsec: 0 opsec: -2147483648
sd57 util: 0% queue: 0 wait: 0 serv: 1 kbsec: 12 opsec: 2
sd40 util: 0% queue: -2147483648 wait: 0 serv: 0 kbsec: 0 opsec: -2147483648
sd44 util: 0% queue: -2147483648 wait: 0 serv: 0 kbsec: 0 opsec: -2147483648
Under Filesystem data - Only this
-
File System Data -
/ capacity: 5120MB free: 3646MB
h: 8 / capacity: 5120MB free: 3646MB
h: 9 / capacity: 5120MB free: 3646MB
Validation Record: VVV
Yes, ST06 --> snapshot displays only root filesystem
Its a zone residing on a unix cluster ( but not a sap cluster, system is a distributed system with CI and DB on seperate zones)
But, we have saposcol running on other zones (similar build) running ECC6 or BI etc, which doesnt have this issue
> Its a zone residing on a unix cluster ( but not a sap cluster, system is a distributed system with CI and DB on seperate zones)
So do you have separate mountpoints? Or is everything residing under /?
What's the output of
cat /etc/vfstab
I can only suggest to upgrade your saposcol to the latest patchlevel and try again. If that does still not work I'd open an OSS call.
Markus
hi, sorry for delay in response
There are seperate mount points..not residing under root
$ cat /etc/vfstab
#device device mount FS fsck mount mount
#to mount to fsck point type pass at boot options
#
/proc - /proc proc - no -
ctfs - /system/contract ctfs - no -
objfs - /system/object objfs - no -
sharefs - /etc/dfs/sharetab sharefs - no -
fd - /dev/fd fd - no -
swap - /tmp tmpfs - yes size=2048M
Its the latest saposcol for 7.1, i took from marketplace
I have opened a OSS call and even they are investigating, but no luck so far
Thanks Markus, excuse for the delay in response
Below is the entry on a similar UNIX zone build having an ECC6 system
This system is showing all local Filesystem but not the NFS ones ( which is not an issue)
$ cat /etc/vfstab
#device device mount FS fsck mount mount
#to mount to fsck point type pass at boot options
#
/proc - /proc proc - no -
ctfs - /system/contract ctfs - no -
objfs - /system/object objfs - no -
sharefs - /etc/dfs/sharetab sharefs - no -
fd - /dev/fd fd - no -
swap - /tmp tmpfs - yes -
swap - /tmp tmpfs - yes size=512M
******005-f05:/vol/ss005_f05_vol2/********** - /sapcd nfs - yes vers=4
********-nfs:/zone/mnt/*********/usr_sap_trans - /usr/sap/trans nfs - yes vers=4
RZ20 --> Operating System -->
Filesystems
/
/dev
/etc/globalzone
/home
/interface
/sapmnt/??? (ecc6 system)
/usr/sap
/var
/lib/libc.so.1
The ones where we are having the problem is (only root) is all having PI 7.1 systems on it
Not sure, whats the problem. Still SAP hasnt come back with any solution
>
> $ cat /etc/vfstab
> #device device mount FS fsck mount mount
> #to mount to fsck point type pass at boot options
> #
> /proc - /proc proc - no -
> ctfs - /system/contract ctfs - no -
> objfs - /system/object objfs - no -
> sharefs - /etc/dfs/sharetab sharefs - no -
> fd - /dev/fd fd - no -
> swap - /tmp tmpfs - yes -
> swap - /tmp tmpfs - yes size=512M
> ******005-f05:/vol/ss005_f05_vol2/********** - /sapcd nfs - yes vers=4
> ********-nfs:/zone/mnt/*********/usr_sap_trans - /usr/sap/trans nfs - yes vers=4
>
So what filesystems did you expect to be seen by saposcol? As you can see yourself, /etc/fstab does not have a /usr/sap or others...
Markus
> This is the saposcol patch which is deployed - latest one available in marketplace
>
> 7.10 - 64bit UNICODE - solaris X86_64
>
> saposcol_30-10004222.sar - patch number 30
Ok - then I'm out of ideas...
You may try, as a test, install the 7.00 saposcol on that system and try if you see them there.
Markus
Hi Shanmuga,
We are experiencing the exact same problem. Did SAP find an answer to the problem. We'd love to know the answer too!
Hi Jeffrey,
This is actually a bug with saposcol on 7.1 kernel
The OSS message i raised with SAP had SUN guys checking it and they asked me to use 7.0 saposcol as a workaround
They confirmed the bug with 7.1 saposcol
So, try using 7.0 saposcol for 7.1 systems and it displays all filesystems correctly
Hi Markus,
The issue is now fixed. the saposcol with 7.1 is an older version with this bug (SUN folks confirmed it). the 7.0 ones in marketplace are newer with the fix applied for this bug
So i deployed 7.00 64bit saposcol and this published all filesystems
Many Thanks again for your quick and helpful replies
regards
Shanmuga
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
> The issue is now fixed. the saposcol with 7.1 is an older version with this bug (SUN folks confirmed it). the 7.0 ones in marketplace are newer with the fix applied for this bug
>
> So i deployed 7.00 64bit saposcol and this published all filesystems
>
> Many Thanks again for your quick and helpful replies
You're welcome!
thank you for the feedback!
Markus
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.