cancel
Showing results for 
Search instead for 
Did you mean: 

NFS issue for SAP

pa8700
Regular Advisor

NFS issue for SAP

We are facing problems in a batch job that runs on our SAP R/3 PRD system.
I found in the SAP system log the following message:

----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
4:22:27 BTC 44 120 BASISBTC F6E Error synchronizing file /usr/sap/PLA/SYS/global/120JOBLG/0001X12540801X03800. Error 70
There are processing modes for which TemSe objects are stored in files and for which, before each COMMIT, it is ensured that data is also written to the physical hard disk (and not only into some cache). In this case, the error which occurred can be examined with the help of the "errno".
----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------

The error occurred twice: first on Jan 25 at 16:22 and the second on Feb 01 at 14:48.

Looking at the file errno.h I hve found the following information
# define ESTALE 70 /* Stale NFS file handle */

I contacted the SAP Support to get help, and the answer is below:

--------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
The error you are getting is error 70 which is a stale NFS file handle.

This error, (stale NFS file handle) is caused by a NFS client request ona file (ie. read/write) using a handle to that file which is no longer
valid
4 REPLIES
pa8700
Regular Advisor

Re: NFS issue for SAP

pls find the nfstat output

# nfsstat -c

Client rpc:
Connection oriented:
calls badcalls badxids
989997 6 4
timeouts newcreds badverfs
6 0 0
timers cantconn nomem
0 0 0
interrupts
0
Connectionless oriented:
calls badcalls retrans
7522 0 144
badxids timeouts waits
0 144 0
newcreds badverfs timers
0 0 0
toobig nomem cantsend
0 0 0
bufulocks
0

Client nfs:
calls badcalls clgets
989991 0 989991
cltoomany
4
Version 2: (0 calls)
null getattr setattr
0 0% 0 0% 0 0%
root lookup readlink
0 0% 0 0% 0 0%
read wrcache write
0 0% 0 0% 0 0%
create remove rename
0 0% 0 0% 0 0%
link symlink mkdir
0 0% 0 0% 0 0%
rmdir readdir statfs
0 0% 0 0% 0 0%
Version 3: (989991 calls)
null getattr setattr
0 0% 3209 0% 34 0%
lookup access readlink
11392 1% 1161 0% 0 0%
read write create
12306 1% 789359 79% 18 0%
mkdir symlink mknod
2 0% 4 0% 0 0%
remove rmdir rename
6 0% 0 0% 0 0%
link readdir readdir+
0 0% 0 0% 414 0%
fsstat fsinfo pathconf
171340 17% 7 0% 0 0%
commit
739 0%

================================================
g# nfsstat -s

Server rpc:
Connection oriented:
calls badcalls nullrecv
39875167 0 0
badlen xdrcall dupchecks
0 0 2872861
dupreqs
157
Connectionless oriented:
calls badcalls nullrecv
45 0 0
badlen xdrcall dupchecks
0 0 0
dupreqs
0

Server nfs:
calls badcalls
39875211 0
Version 2: (0 calls)
null getattr setattr
0 0% 0 0% 0 0%
root lookup readlink
0 0% 0 0% 0 0%
read wrcache write
0 0% 0 0% 0 0%
create remove rename
0 0% 0 0% 0 0%
link symlink mkdir
0 0% 0 0% 0 0%
rmdir readdir statfs
0 0% 0 0% 0 0%
Version 3: (39875197 calls)
null getattr setattr
3776 0% 953241 2% 43787 0%
lookup access readlink
6576070 16% 14875726 37% 3 0%
read write create
10228131 25% 2635634 6% 121164 0%
mkdir symlink mknod
0 0% 0 0% 0 0%
remove rmdir rename
72261 0% 0 0% 1 0%
link readdir readdir+
0 0% 30 0% 275214 0%
fsstat fsinfo pathconf
1963635 4% 3750 0% 0 0%
commit
2122774 5%

pa8700
Regular Advisor

Re: NFS issue for SAP

I have some errors in the syslog as below

automountd[1042]: code : RPC: Unknown host
Michael Steele_2
Honored Contributor

Re: NFS issue for SAP

Hi

The error occurred twice: first on Jan 25 at 16:22 and the second on Feb 01 at 14:48.

What else was running at the end of the day / afternoon just before everybody in the company goes home?

Capture the process list next time you run and you will probably see something big contending for the same data.
Support Fatherhood - Stop Family Law
Dennis Handly
Acclaimed Contributor

Re: NFS issue for SAP

>Looking at the file errno.h

I would feel a lot safer if the error message said "errno 70" (or even ESTALE) and not just "Error". Anyone can create their own error ranges. But SAP Support confirms that correspondence.