[LU-1668] Test failure on test suite conf-sanity 53a Created: 24/Jul/12 Updated: 16/Feb/14 Resolved: 16/Feb/14 |
|
| Status: | Resolved |
| Project: | Lustre |
| Component/s: | None |
| Affects Version/s: | Lustre 2.3.0 |
| Fix Version/s: | None |
| Type: | Bug | Priority: | Minor |
| Reporter: | Maloo | Assignee: | Liang Zhen (Inactive) |
| Resolution: | Fixed | Votes: | 0 |
| Labels: | None | ||
| Environment: |
lustre-master-tag-2.2.91 OFED build |
||
| Severity: | 3 |
| Rank (Obsolete): | 4486 |
| Description |
|
This issue was created by maloo for sarah <sarah@whamcloud.com> This issue relates to the following test suite run: https://maloo.whamcloud.com/test_sets/5b019fd0-d5c0-11e1-b078-52540035b04c. == conf-sanity test 53a: check OSS thread count params == 13:54:59 (1343076899) Loading modules from /usr/lib64/lustre detected 4 online CPUs by sysfs Force libcfs to create 2 CPU partitions ../libcfs/libcfs/libcfs options: 'cpu_npartitions=2' WARNING: Deprecated config file /etc/modprobe.conf, all config files belong into /etc/modprobe.d/. debug=-1 subsystem_debug=0xffb7efff ../lnet/lnet/lnet options: 'accept=all networks="o2ib0(ib0),tcp0(eth0)" accept_port=7988' WARNING: Deprecated config file /etc/modprobe.conf, all config files belong into /etc/modprobe.d/. WARNING: Deprecated config file /etc/modprobe.conf, all config files belong into /etc/modprobe.d/. WARNING: Deprecated config file /etc/modprobe.conf, all config files belong into /etc/modprobe.d/. WARNING: Deprecated config file /etc/modprobe.conf, all config files belong into /etc/modprobe.d/. WARNING: Deprecated config file /etc/modprobe.conf, all config files belong into /etc/modprobe.d/. gss/krb5 is not supported WARNING: Deprecated config file /etc/modprobe.conf, all config files belong into /etc/modprobe.d/. WARNING: Deprecated config file /etc/modprobe.conf, all config files belong into /etc/modprobe.d/. WARNING: Deprecated config file /etc/modprobe.conf, all config files belong into /etc/modprobe.d/. WARNING: Deprecated config file /etc/modprobe.conf, all config files belong into /etc/modprobe.d/. WARNING: Deprecated config file /etc/modprobe.conf, all config files belong into /etc/modprobe.d/. WARNING: Deprecated config file /etc/modprobe.conf, all config files belong into /etc/modprobe.d/. WARNING: Deprecated config file /etc/modprobe.conf, all config files belong into /etc/modprobe.d/. WARNING: Deprecated config file /etc/modprobe.conf, all config files belong into /etc/modprobe.d/. WARNING: Deprecated config file /etc/modprobe.conf, all config files belong into /etc/modprobe.d/. start mds service on client-3-ib Starting mds1: -o user_xattr,acl /dev/sda3 /mnt/mds1 Started lustre-MDT0000 start ost1 service on fat-amd-4-ib Starting ost1: /dev/sdb1 /mnt/ost1 Started lustre-OST0000 mount lustre on /mnt/lustre..... Starting client: client-4.lab.whamcloud.com: -o flock,user_xattr,acl client-3-ib@o2ib:/lustre /mnt/lustre setup single mount lustre success checking (($tstarted && $tmin && $tmax)) (((14 && 12 && 384)))... checking (($tstarted >= $tmin && $tstarted <= $tmax )) (((14 >= 12 && 14 <= 384 )))... ost.OSS.ost.threads_min=14 ost.OSS.ost.threads_max=382 checking (($tmin2 == ($tmin + $ncpts) && $tmax2 == ($tmax - $ncpts))) (((12 == (12 + 2) && 380 == (384 - 2))))... conf-sanity test_53a: @@@@@@ FAIL: Assertion 25 failed: (($tmin2 == ($tmin + $ncpts) && $tmax2 == ($tmax - $ncpts))) (expanded: ((12 == (12 + 2) && 380 == (384 - 2)))) Insane OST thread counts |
| Comments |
| Comment by Liang Zhen (Inactive) [ 10/Aug/12 ] |
thread_sanity() {
local modname=$1
local facet=$2
local parampat=$3
local opts=$4
local tmin
local tmin2
local tmax
local tmax2
local tstarted
local paramp
local msg="Insane $modname thread counts"
local ncpts=$(check_cpt_number)
check_cpt_number is checking ncpts on localhost, but it really should be "do_fact $facet check_cpt_number", I will post a patch soon |
| Comment by Liang Zhen (Inactive) [ 10/Aug/12 ] |
|
Patch is here: http://review.whamcloud.com/3595 |
| Comment by James A Simmons [ 10/Aug/12 ] |
|
My testing also has run into this bug. I'm still failing with the patch. 53b (Side note - I'm running with 1024 mds threads by default) setup single mount lustre success For 53a I get mount lustre on /lustre/barry..... |
| Comment by Sarah Liu [ 14/Aug/12 ] |
|
another failure, client is running kernel 2.6.38-fc15 https://maloo.whamcloud.com/test_sets/b0206b50-e5c3-11e1-ae4e-52540035b04c |
| Comment by James A Simmons [ 16/Aug/12 ] |
|
Looking at the data for 53a I'm getting tmin = 6 Which causes the ASSERT to fail |
| Comment by James A Simmons [ 17/Aug/12 ] |
|
I found the reason for the failure for my testing. In /etc/modprobe.d/lustre.conf I have options mdt mds_num_threads=1024 When I pound it out both test pass. |
| Comment by Liang Zhen (Inactive) [ 17/Aug/12 ] |
|
hi James, then could you change the review flag? or you want to me to make more change to the patch so it can pass even with setting this parameter? I think this parameter is conflicting with the conf-sanity because conf-sanity needs to set this parameter as well. |
| Comment by James A Simmons [ 17/Aug/12 ] |
|
Yes could you arrange for the test to pass even in the case with setting this parameter. |
| Comment by James A Simmons [ 17/Aug/12 ] |
|
Okay I can change my review since your current patch cover most cases whereas my problem is a specific corner case. If I can change my review can you keep this ticket open until my problem is resolved? |
| Comment by Peter Jones [ 20/Aug/12 ] |
|
Dropping priority as patch covering most cases has been landed to master for 2.3 |
| Comment by Sarah Liu [ 26/Dec/12 ] |
|
another failure seen in interop testing between 2.1.3 server and 2.4 client: https://maloo.whamcloud.com/test_sets/351a6028-4a86-11e2-8a7b-52540035b04c |
| Comment by Jian Yu [ 19/Dec/13 ] |
|
Lustre client: http://build.whamcloud.com/job/lustre-b2_3/41/ (2.3.0) conf-sanity test 53b failed: |
| Comment by Liang Zhen (Inactive) [ 16/Feb/14 ] |
|
I think this has already been fixed |