Project

General

Profile

Actions

action #64707

closed

[qac][public cloud][ltp] epoll_wait02 test from LTP syscalls fails on GCE On-Demand

Added by jlausuch about 4 years ago. Updated about 4 years ago.

Status:
Resolved
Priority:
Normal
Assignee:
-
Category:
Bugs in existing tests
Target version:
-
Start date:
2020-03-20
Due date:
% Done:

0%

Estimated time:
Difficulty:

Description

https://openqa.suse.de/tests/4019168#step/epoll_wait02/1

However, on other images it passes, e.g. https://openqa.suse.de/tests/4017274#step/epoll_wait02/1

This is the output form the test execution:

sh-4.4# epoll_wait02; echo cmd-exit-129-$?
tst_test.c:1229: INFO: Timeout per run is 0h 05m 00s
tst_timer_test.c:348: INFO: CLOCK_MONOTONIC resolution 1ns
tst_timer_test.c:360: INFO: prctl(PR_GET_TIMERSLACK) = 50us
tst_timer_test.c:264: INFO: epoll_wait() sleeping for 1000us 500 iterations, threshold 450.01us
tst_timer_test.c:307: INFO: min 1068us, max 1167us, median 1073us, trunc mean 1074.28us (discarded 25)
tst_timer_test.c:322: PASS: Measured times are within thresholds
tst_timer_test.c:264: INFO: epoll_wait() sleeping for 2000us 500 iterations, threshold 450.01us
tst_timer_test.c:307: INFO: min 2071us, max 2174us, median 2077us, trunc mean 2078.58us (discarded 25)
tst_timer_test.c:322: PASS: Measured times are within thresholds
tst_timer_test.c:264: INFO: epoll_wait() sleeping for 5000us 300 iterations, threshold 450.04us
tst_timer_test.c:307: INFO: min 5075us, max 5209us, median 5084us, trunc mean 5087.58us (discarded 15)
tst_timer_test.c:322: PASS: Measured times are within thresholds
tst_timer_test.c:264: INFO: epoll_wait() sleeping for 10000us 100 iterations, threshold 450.33us
tst_timer_test.c:307: INFO: min 10080us, max 10167us, median 10091us, trunc mean 10092.68us (discarded 5)
tst_timer_test.c:322: PASS: Measured times are within thresholds
tst_timer_test.c:264: INFO: epoll_wait() sleeping for 25000us 50 iterations, threshold 451.29us
tst_timer_test.c:307: INFO: min 25085us, max 25187us, median 25106us, trunc mean 25107.94us (discarded 2)
tst_timer_test.c:322: PASS: Measured times are within thresholds
tst_timer_test.c:264: INFO: epoll_wait() sleeping for 100000us 10 iterations, threshold 537.00us
tst_timer_test.c:307: INFO: min 100154us, max 104057us, median 100165us, trunc mean 100616.56us (discarded 1)
tst_timer_test.c:310: FAIL: epoll_wait() slept for too long
Time: us | Frequency
--------------------------------------------------------------------------------
100154 | ********************************************************************
100360 | *********+
100566 | 
100772 | 
100978 | 
101184 | 
101390 | 
101596 | 
101802 | 
102008 | 
102214 | 
102420 | 
102626 | 
102832 | 
103038 | 
103244 | 
103450 | 
103656 | 
103862 | *******************-
--------------------------------------------------------------------------------
206us | 1 sample = 9.71429 '*', 19.42857 '+', 38.85714 '-', non-zero '.'
tst_timer_test.c:264: INFO: epoll_wait() sleeping for 1000000us 2 iterations, threshold 4400.00us
tst_timer_test.c:307: INFO: min 1037244us, max 1039297us, median 1037244us, trunc mean 1037244.00us (discarded 1)
tst_timer_test.c:310: FAIL: epoll_wait() slept for too long
Time: us | Frequency
--------------------------------------------------------------------------------
1037244 | ********************************************************************
1037353 | 
1037462 | 
1037571 | 
1037680 | 
1037789 | 
1037898 | 
1038007 | 
1038116 | 
1038225 | 
1038334 | 
1038443 | 
1038552 | 
1038661 | 
1038770 | 
1038879 | 
1038988 | 
1039097 | 
1039206 | ********************************************************************
--------------------------------------------------------------------------------
109us | 1 sample = 68.00000 '*', 136.00000 '+', 272.00000 '-', non-zero '.'
Summary:
passed   5
failed   2
skipped  0
warnings 0    
Actions #1

Updated by jlausuch about 4 years ago

  • Category set to Bugs in existing tests
  • Priority changed from Normal to High
  • Target version set to 445
Actions #2

Updated by cfconrad about 4 years ago

Running it manually on an GCE n1-standard-8 instance, this error doesn't appear. Don't know why we see this error when using the TA.
Maybe it is related to some other test we run before.

clemix@cfconrad-epoll-wait02:/opt/ltp/testcases/bin> ./epoll_wait02 
tst_test.c:1229: INFO: Timeout per run is 0h 05m 00s
tst_timer_test.c:348: INFO: CLOCK_MONOTONIC resolution 1ns
tst_timer_test.c:360: INFO: prctl(PR_GET_TIMERSLACK) = 50us
tst_timer_test.c:370: INFO: Failed to set zero latency constraint: Permission denied
tst_timer_test.c:264: INFO: epoll_wait() sleeping for 1000us 500 iterations, threshold 450.01us
tst_timer_test.c:307: INFO: min 1043us, max 1137us, median 1081us, trunc mean 1084.77us (discarded 25)
tst_timer_test.c:322: PASS: Measured times are within thresholds
tst_timer_test.c:264: INFO: epoll_wait() sleeping for 2000us 500 iterations, threshold 450.01us
tst_timer_test.c:307: INFO: min 2036us, max 2134us, median 2082us, trunc mean 2084.95us (discarded 25)
tst_timer_test.c:322: PASS: Measured times are within thresholds
tst_timer_test.c:264: INFO: epoll_wait() sleeping for 5000us 300 iterations, threshold 450.04us
tst_timer_test.c:307: INFO: min 5076us, max 5141us, median 5087us, trunc mean 5089.39us (discarded 15)
tst_timer_test.c:322: PASS: Measured times are within thresholds
tst_timer_test.c:264: INFO: epoll_wait() sleeping for 10000us 100 iterations, threshold 450.33us
tst_timer_test.c:307: INFO: min 10079us, max 10131us, median 10092us, trunc mean 10094.33us (discarded 5)
tst_timer_test.c:322: PASS: Measured times are within thresholds
tst_timer_test.c:264: INFO: epoll_wait() sleeping for 25000us 50 iterations, threshold 451.29us
tst_timer_test.c:307: INFO: min 25086us, max 25356us, median 25098us, trunc mean 25099.96us (discarded 2)
tst_timer_test.c:322: PASS: Measured times are within thresholds
tst_timer_test.c:264: INFO: epoll_wait() sleeping for 100000us 10 iterations, threshold 537.00us
tst_timer_test.c:307: INFO: min 100151us, max 100199us, median 100165us, trunc mean 100163.56us (discarded 1)
tst_timer_test.c:322: PASS: Measured times are within thresholds
tst_timer_test.c:264: INFO: epoll_wait() sleeping for 1000000us 2 iterations, threshold 4400.00us
tst_timer_test.c:307: INFO: min 1001045us, max 1001051us, median 1001045us, trunc mean 1001045.00us (discarded 1)
tst_timer_test.c:322: PASS: Measured times are within thresholds
Summary:
passed   7
failed   0
skipped  0
warnings 0
Actions #3

Updated by cfconrad about 4 years ago

I also retried with run_ltp, as we do in openqa.

clemix@linux-7k5u:~/gce/ltp_epoll_wait02/2020-03-22> perl -I ltp/tools/runltp-ng ltp/tools/runltp-ng/runltp-ng --logname=ltp_log --run syscalls --exclude '_16$|^(move_pages|msg(ctl|get|rcv|snd|stress)|sem(ctl|get|op|snd)|shm(at|ctl|dt|get)|syslog|inotify07|inotify08)' --backend=ssh:user=clemix:key_file=/home/clemix/id_rsa:host=34.89.134.250

But still don't get that error :/

Actions #4

Updated by jlausuch about 4 years ago

  • Subject changed from [kernel][public cloud][ltp] epoll_wait02 test from LTP syscalls fails on GCE On-Demand to [qac][public cloud][ltp] epoll_wait02 test from LTP syscalls fails on GCE On-Demand
Actions #5

Updated by jlausuch about 4 years ago

  • Target version deleted (445)
Actions #7

Updated by MDoucha about 4 years ago

This appears to be just another case of randomly overloaded VM host.

Actions #8

Updated by jlausuch about 4 years ago

  • Priority changed from High to Normal

Ok. Lowering the prio of this ticket as it's not critical and sporadic depending on hypervisor. Let's keep it open in case it appears again.

Actions #9

Updated by cfconrad about 4 years ago

  • Status changed from New to Resolved

Excluded with:

COMMAND_EXCLUDE=_16$|^(move_pages|msg(ctl|get|rcv|snd|stress)|sem(ctl|get|op|snd)|shm(at|ctl|dt|get)|syslog|inotify07|inotify08|epoll_wait02)

As discussed, this test is not reliable for VMs

Actions

Also available in: Atom PDF