[#LU-1249] lctl set_param debug_mb=10 gives Invalid argument

[LU-1249] lctl set_param debug_mb=10 gives Invalid argument Created: 22/Mar/12 Updated: 26/Aug/12 Resolved: 30/Apr/12
Status:	Resolved
Project:	Lustre
Component/s:	None
Affects Version/s:	Lustre 2.3.0, Lustre 2.1.1
Fix Version/s:	Lustre 2.3.0, Lustre 2.1.3

Type:

Bug

Priority:

Blocker

Reporter:

Frey Alfredsson

Assignee:

Zhenyu Xu

Resolution:

Fixed

Votes:

Labels:

None

Environment:

RHEL 6.2 with no updates.
Installed using Platform HPC 3.1 (build 7211) and does therefore contain extra packages from Platform Computing.

Issue Links:

Duplicate
is duplicated by	~~LU-958~~	tests: debug_mb set incorrectly for s...	Resolved

Severity:

Rank (Obsolete):

4639

Description

Step 1: Install RHEL 6.2 using Platform HPC 3.1.
Apart from the Platform HPC 3.1 extra packages, there are no custom changes to the RHEl 6.2 distribution nor updates at this point.

Step 2: Packages installed.
kernel-2.6.32-220.el6_lustre.g4554b65.x86_64.rpm
kernel-debuginfo-2.6.32-220.el6_lustre.g4554b65.x86_64.rpm
kernel-debuginfo-common-x86_64-2.6.32-220.el6_lustre.g4554b65.x86_64.rpm
kernel-devel-2.6.32-220.el6_lustre.g4554b65.x86_64.rpm
kernel-firmware-2.6.32-220.el6_lustre.g4554b65.x86_64.rpm
kernel-headers-2.6.32-220.el6_lustre.g4554b65.x86_64.rpm
lustre-2.1.1-2.6.32_220.el6_lustre.g4554b65.x86_64.x86_64.rpm
lustre-debuginfo-2.1.1-2.6.32_220.el6_lustre.g4554b65.x86_64.x86_64.rpm
lustre-ldiskfs-3.3.0-2.6.32_220.el6_lustre.g4554b65.x86_64.x86_64.rpm
lustre-ldiskfs-debuginfo-3.3.0-2.6.32_220.el6_lustre.g4554b65.x86_64.x86_64.rpm
lustre-modules-2.1.1-2.6.32_220.el6_lustre.g4554b65.x86_64.x86_64.rpm
lustre-source-2.1.1-2.6.32_220.el6_lustre.g4554b65.x86_64.x86_64.rpm
lustre-tests-2.1.1-2.6.32_220.el6_lustre.g4554b65.x86_64.x86_64.rpm
perf-2.6.32-220.el6_lustre.g4554b65.x86_64.rpm
perf-debuginfo-2.6.32-220.el6_lustre.g4554b65.x86_64.rpm
python-perf-2.6.32-220.el6_lustre.g4554b65.x86_64.rpm

Step 3: Computer rebooted.

Step 4: Run lctl set_param debug_mb=10
[root@storage000 ~]# lctl set_param debug_mb=10
debug_mb=10
error: set_param: writing to file /proc/sys/lnet/debug_mb: Invalid argument
[root@storage000 ~]#

[root@storage000 ~]# cat /proc/sys/lnet/debug_mb
121
[root@storage000 ~]# echo 10 > /proc/sys/lnet/debug_mb
-bash: echo: write error: Invalid argument
[root@storage000 ~]#

This error is reproduced every time after following the steps above.

Comments

Comment by Frey Alfredsson [ 22/Mar/12 ]

It seems that I need to have lctl set_param debug_mb=24 for this to work.

When I ran /usr/lib64/lustre/tests/llmount.sh it automatically tried to use debug_mb=10, that is why it became a problem.

Comment by Frey Alfredsson [ 22/Mar/12 ]

This machine has one E5603 CPU with 4 cores.

Comment by Johann Lombardi (Inactive) [ 22/Mar/12 ]

I think the llmount issue has been fixed in this ticket in ~~LU-958~~

Comment by Andreas Dilger [ 22/Mar/12 ]

The llmount fix is only really in the test scripts and do not fix this problem for users at all. I think that we should make the proc tunable more forgiving - arriving it to the minimal value if a value lower than this is used, and the maximum if a value higher than this is used.

At that point it makes sense to also remove the workarounds in the test scripts.

Comment by Peter Jones [ 06/Apr/12 ]

Bobi

Could you please look at this one?

Thanks

Peter

Comment by Zhenyu Xu [ 09/Apr/12 ]

patch tracking at http://review.whamcloud.com/2489