Skip to contentSkip to conversation

tests: enable meson test timeouts to improve debuggability

PATCH In review Daniel P. Berrangé 5 participants 29 messages Jul 17, 2023 – Sep 12, 2023 Arm/Amlogic Meson SoC support

lore
Share & tools
lore
#0 lore raw

v2 cover tests: enable meson test timeouts to improve debuggability

Perhaps the most painful of all the GitLab CI failures we see are the enforced job timeouts:

"ERROR: Job failed: execution took longer than 1h15m0s seconds"

https://gitlab.com/qemu-project/qemu/-/jobs/4387047648

when that hits the CI log shows what has *already* run, but figuring out what was currently running (or rather stuck) is an horrendously difficult.

The initial meson port disabled the meson test timeouts, in order to limit the scope for introducing side effects from the port that would complicate adoption.

Now that the meson port is basically finished we can take advantage of more of its improved features. It has the ability to set timeouts for test programs, defaulting to 30 seconds, but overridable per test. This is further helped by fact that we changed the iotests integration so that each iotests was a distinct meson test, instead of having one single giant (slow) test.

We already set overrides for a bunch of tests, but they've not been kept up2date since we had timeouts disabled. So this series first updates the timeout overrides such that all tests pass when run in my test gitlab CI pipeline. Then it enables use of meson timeouts.

We might still hit timeouts due to non-deterministic performance of gitlab CI runners. So we'll probably have to increase a few more timeouts in the short term. Fortunately this is going to be massively easier to diagnose. For example this job during my testing:

https://gitlab.com/berrange/qemu/-/jobs/4392029495

we can immediately see the problem tests

Summary of Failures:

  6/252 qemu:qtest+qtest-i386 / qtest-i386/bios-tables-test                TIMEOUT        120.02s   killed by signal 15 SIGTERM
  7/252 qemu:qtest+qtest-aarch64 / qtest-aarch64/bios-tables-test          TIMEOUT        120.03s   killed by signal 15 SIGTERM
 64/252 qemu:qtest+qtest-aarch64 / qtest-aarch64/qom-test                  TIMEOUT        300.03s   killed by signal 15 SIGTERM

The full meson testlog.txt will show each individual TAP log output, so we can then see exactly which test case we got stuck on.

NB, the artifacts are missing on the job links above, until this patch merges:

https://lists.gnu.org/archive/html/qemu-devel/2023-05/msg04668.html

Changed in v2:

* Increase timeouts for many more tests, such that an --enable-debug build stands a better chance of passing tests too, without the user manually setting a timeout multiplier for meson.

Daniel P. Berrangé (12):
qtest: bump min meson timeout to 60 seconds
qtest: bump migration-test timeout to 5 minutes
qtest: bump qom-test timeout to 15 minutes
qtest: bump npcm7xx_pwn-test timeout to 5 minutes
qtest: bump test-hmp timeout to 4 minutes
qtest: bump pxe-test timeout to 3 minutes
qtest: bump prom-env-test timeout to 3 minutes
qtest: bump boot-serial-test timeout to 3 minutes
qtest: bump qos-test timeout to 2 minutes
qtest: bump aspeed_smc-test timeout to 4 minutes
qtest: bump bios-table-test timeout to 9 minutes
mtest2make: stop disabling meson test timeouts

 scripts/mtest2make.py   |  3 ++-
 tests/qtest/meson.build | 24 ++++++++++++------------
 2 files changed, 14 insertions(+), 13 deletions(-)
Signature · 1 line
2.41.0
#1 lore raw

PATCH v2 1/12 qtest: bump min meson timeout to 60 seconds

Even some of the relatively fast qtests can sometimes hit the 30 second timeout in GitLab CI under high parallelism/load conditions. Bump the min to 60 seconds to give a higher margin for reliability.

---
 tests/qtest/meson.build | 9 ++-------
 1 file changed, 2 insertions(+), 7 deletions(-)
tests/qtest/meson.build +2 -7
index b071d400b3..b29d7abc39 100644
@@ -1,12 +1,7 @@
 slow_qtests = {
-  'ahci-test' : 60,
   'bios-tables-test' : 120,
-  'boot-serial-test' : 60,
   'migration-test' : 150,
   'npcm7xx_pwm-test': 150,
-  'prom-env-test' : 60,
-  'pxe-test' : 60,
-  'qos-test' : 60,
   'qom-test' : 300,
   'test-hmp' : 120,
 }
@@ -378,8 +373,8 @@ foreach dir : target_dirs
          env: qtest_env,
          args: ['--tap', '-k'],
          protocol: 'tap',
-         timeout: slow_qtests.get(test, 30),
-         priority: slow_qtests.get(test, 30),
+         timeout: slow_qtests.get(test, 60),
+         priority: slow_qtests.get(test, 60),
          suite: ['qtest', 'qtest-' + target_base])
   endforeach
 endforeach
Signature · 1 line
2.41.0
#2 lore raw

PATCH v2 2/12 qtest: bump migration-test timeout to 5 minutes

In reply to Daniel P. Berrangé

The migration test should take between 1 min 30 and 2 mins on reasonably modern hardware. The test is not especially compute bound, rather its running time is dominated by the guest RAM size relative to the bandwidth cap, which forces each iteration to take at least 30 seconds. None the less under high load conditions with multiple QEMU processes spawned and competing with other parallel tests, the worst case running time might be somewhat extended. Bumping the timeout to 5 minutes gives us good headroom, while still catching stuck tests relatively quickly.

---
 tests/qtest/meson.build | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
tests/qtest/meson.build +1 -1
index b29d7abc39..c3f8bf45e4 100644
@@ -1,6 +1,6 @@
 slow_qtests = {
   'bios-tables-test' : 120,
-  'migration-test' : 150,
+  'migration-test' : 300,
   'npcm7xx_pwm-test': 150,
   'qom-test' : 300,
   'test-hmp' : 120,
Signature · 1 line
2.41.0
#3 lore raw

PATCH v2 3/12 qtest: bump qom-test timeout to 15 minutes

In reply to Daniel P. Berrangé

The qom-test is periodically hitting the 5 minute timeout when running on the aarch64 emulator under GitLab CI. With an --enable-debug build it can take over 10 minutes for arm/aarch64 targets. Setting timeout to 15 minutes gives enough headroom to hopefully make it reliable.

---
 tests/qtest/meson.build | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
tests/qtest/meson.build +1 -1
index c3f8bf45e4..ecbf515d50 100644
@@ -2,7 +2,7 @@ slow_qtests = {
   'bios-tables-test' : 120,
   'migration-test' : 300,
   'npcm7xx_pwm-test': 150,
-  'qom-test' : 300,
+  'qom-test' : 900,
   'test-hmp' : 120,
 }
 
Signature · 1 line
2.41.0
#4 lore raw

PATCH v2 4/12 qtest: bump npcm7xx_pwn-test timeout to 5 minutes

In reply to Daniel P. Berrangé

The npcm7xx_pwn-test takes 3 & 1/2 minutes in a --enable-debug build. Bumping to 5 minutes will give more headroom.

---
 tests/qtest/meson.build | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
tests/qtest/meson.build +1 -1
index ecbf515d50..e28f0925e1 100644
@@ -1,7 +1,7 @@
 slow_qtests = {
   'bios-tables-test' : 120,
   'migration-test' : 300,
-  'npcm7xx_pwm-test': 150,
+  'npcm7xx_pwm-test': 300,
   'qom-test' : 900,
   'test-hmp' : 120,
 }
Signature · 1 line
2.41.0
Reviewed-by on v2 4/12
#5 lore raw

Re: [PATCH v2 04/12] qtest: bump npcm7xx_pwn-test timeout to 5 minutes

Quoted 13 lines · Daniel P. Berrangé

The npcm7xx_pwn-test takes 3 & 1/2 minutes in a --enable-debug build. Bumping to 5 minutes will give more headroom.

---
  tests/qtest/meson.build | 2 +-
  1 file changed, 1 insertion(+), 1 deletion(-)
tests/qtest/meson.build +0 -0
index ecbf515d50..e28f0925e1 100644

tests/qtest/meson.build line 7

Quoted 8 lines
tests/qtest/meson.build +1 -1
@@ -1,7 +1,7 @@
  slow_qtests = {
    'bios-tables-test' : 120,
    'migration-test' : 300,
-  'npcm7xx_pwm-test': 150,
+  'npcm7xx_pwm-test': 300,
    'qom-test' : 900,
    'test-hmp' : 120,
  }

(we should maybe still have a look at this test later to see why it is so slow, though)

#6 lore raw

Re: [PATCH v2 04/12] qtest: bump npcm7xx_pwn-test timeout to 5 minutes

Quoted 26 lines · Thomas Huth
Quoted 13 lines · Daniel P. Berrangé

The npcm7xx_pwn-test takes 3 & 1/2 minutes in a --enable-debug build. Bumping to 5 minutes will give more headroom.

---
  tests/qtest/meson.build | 2 +-
  1 file changed, 1 insertion(+), 1 deletion(-)
tests/qtest/meson.build +0 -0
index ecbf515d50..e28f0925e1 100644

tests/qtest/meson.build line 7

Quoted 8 lines
tests/qtest/meson.build +1 -1
@@ -1,7 +1,7 @@
  slow_qtests = {
    'bios-tables-test' : 120,
    'migration-test' : 300,
-  'npcm7xx_pwm-test': 150,
+  'npcm7xx_pwm-test': 300,
    'qom-test' : 900,
    'test-hmp' : 120,
  }
Quoted 2 lines

(we should maybe still have a look at this test later to see why it is so slow, though)

It seems to be doing a massively OTT amount of work to test a single device. Nuvoton maintainers, can you have a look at cutting it down, please?

thanks
-- PMM

#7 lore raw

PATCH v2 5/12 qtest: bump test-hmp timeout to 4 minutes

In reply to Daniel P. Berrangé

The npcm7xx_pwn-test takes just under 3 minutes in a --enable-debug build. Bumping to 4 minutes will give more headroom.

---
 tests/qtest/meson.build | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
tests/qtest/meson.build +1 -1
index e28f0925e1..1070c6dc64 100644
@@ -3,7 +3,7 @@ slow_qtests = {
   'migration-test' : 300,
   'npcm7xx_pwm-test': 300,
   'qom-test' : 900,
-  'test-hmp' : 120,
+  'test-hmp' : 240,
 }
 
 qtests_generic = [
Signature · 1 line
2.41.0
Reviewed-by on v2 5/12
#8 lore raw

Re: [PATCH v2 05/12] qtest: bump test-hmp timeout to 4 minutes

Quoted 1 line · Daniel P. Berrangé

The npcm7xx_pwn-test takes just under 3 minutes in a --enable-debug

s/npcm7xx_pwn-test/test-hmp/

With that copy-n-paste error fixed:

Quoted 20 lines

build. Bumping to 4 minutes will give more headroom.

---
  tests/qtest/meson.build | 2 +-
  1 file changed, 1 insertion(+), 1 deletion(-)
tests/qtest/meson.build +1 -1
index e28f0925e1..1070c6dc64 100644
@@ -3,7 +3,7 @@ slow_qtests = {
    'migration-test' : 300,
    'npcm7xx_pwm-test': 300,
    'qom-test' : 900,
-  'test-hmp' : 120,
+  'test-hmp' : 240,
  }
 
  qtests_generic = [
#9 lore raw

PATCH v2 6/12 qtest: bump pxe-test timeout to 3 minutes

In reply to Daniel P. Berrangé

The pxe-test takes about 1 + 1/2 minutes in a --enable-debug build. Bumping to 3 minutes will give more headroom.

---
 tests/qtest/meson.build | 1 +
 1 file changed, 1 insertion(+)
tests/qtest/meson.build +1 -0
index 1070c6dc64..c6da428dc5 100644
@@ -4,6 +4,7 @@ slow_qtests = {
   'npcm7xx_pwm-test': 300,
   'qom-test' : 900,
   'test-hmp' : 240,
+  'pxe-test': 180,
 }
 
 qtests_generic = [
Signature · 1 line
2.41.0
Reviewed-by on v2 6/12
#10 lore raw

Re: [PATCH v2 06/12] qtest: bump pxe-test timeout to 3 minutes

Quoted 13 lines · Daniel P. Berrangé

The pxe-test takes about 1 + 1/2 minutes in a --enable-debug build. Bumping to 3 minutes will give more headroom.

---
  tests/qtest/meson.build | 1 +
  1 file changed, 1 insertion(+)
tests/qtest/meson.build +0 -0
index 1070c6dc64..c6da428dc5 100644

tests/qtest/meson.build slow_qtests = { · line 10

Quoted 7 lines
tests/qtest/meson.build +1 -0
@@ -4,6 +4,7 @@ slow_qtests = {
    'npcm7xx_pwm-test': 300,
    'qom-test' : 900,
    'test-hmp' : 240,
+  'pxe-test': 180,
  }
 
  qtests_generic = [
#11 lore raw

Re: [PATCH v2 06/12] qtest: bump pxe-test timeout to 3 minutes

In reply to Daniel P. Berrangé
Quoted 13 lines · Daniel P. Berrangé

The pxe-test takes about 1 + 1/2 minutes in a --enable-debug build. Bumping to 3 minutes will give more headroom.

---
  tests/qtest/meson.build | 1 +
  1 file changed, 1 insertion(+)
tests/qtest/meson.build +0 -0
index 1070c6dc64..c6da428dc5 100644

tests/qtest/meson.build slow_qtests = { · line 8

Quoted 5 lines
tests/qtest/meson.build +1 -0
@@ -4,6 +4,7 @@ slow_qtests = {
    'npcm7xx_pwm-test': 300,
    'qom-test' : 900,
    'test-hmp' : 240,
+  'pxe-test': 180,
  }

By the way, pxe-test uses the boot_sector_test() function, and that already uses a timeout of 600 seconds ... so we maybe should use the same (or even slightly higher) timeout value here?

Thomas

#12 lore raw

Re: [PATCH v2 06/12] qtest: bump pxe-test timeout to 3 minutes

Quoted 9 lines · Thomas Huth
Quoted 8 lines · Daniel P. Berrangé

The pxe-test takes about 1 + 1/2 minutes in a --enable-debug build. Bumping to 3 minutes will give more headroom.

---
  tests/qtest/meson.build | 1 +
  1 file changed, 1 insertion(+)
Quoted 14 lines

tests/qtest/meson.build slow_qtests = { · line 8

Quoted 10 lines
tests/qtest/meson.build +1 -0
index 1070c6dc64..c6da428dc5 100644
@@ -4,6 +4,7 @@ slow_qtests = {
    'npcm7xx_pwm-test': 300,
    'qom-test' : 900,
    'test-hmp' : 240,
+  'pxe-test': 180,
  }

By the way, pxe-test uses the boot_sector_test() function, and that already uses a timeout of 600 seconds ... so we maybe should use the same (or even slightly higher) timeout value here?

Your original change setting that 600 limit says:

If QEMU has been compiled with the flags --enable-tcg-interpreter and --enable-debug, the guest is running incredibly slow. The pxe boot test can take up to 400 seconds when testing the pseries ppc64 machine.

so yeah, we need to set the timeout for pxe-test to 600 seconds too. I guess likewise for other tests using boot_sector_test() function too.

With regards,
Daniel

Signature · 3 lines
|: https://berrange.com      -o-    https://www.flickr.com/photos/dberrange :|
|: https://libvirt.org         -o-            https://fstop138.berrange.com :|
|: https://entangle-photo.org    -o-    https://www.instagram.com/dberrange :|
#13 lore raw

PATCH v2 7/12 qtest: bump prom-env-test timeout to 3 minutes

In reply to Daniel P. Berrangé

The prom-env-test takes about 1 + 1/2 minutes in a --enable-debug build. Bumping to 3 minutes will give more headroom.

---
 tests/qtest/meson.build | 1 +
 1 file changed, 1 insertion(+)
tests/qtest/meson.build +1 -0
index c6da428dc5..095c98820e 100644
@@ -5,6 +5,7 @@ slow_qtests = {
   'qom-test' : 900,
   'test-hmp' : 240,
   'pxe-test': 180,
+  'prom-env-test': 180,
 }
 
 qtests_generic = [
Signature · 1 line
2.41.0
#14 lore raw

Re: [PATCH v2 07/12] qtest: bump prom-env-test timeout to 3 minutes

Quoted 13 lines · Daniel P. Berrangé

The prom-env-test takes about 1 + 1/2 minutes in a --enable-debug build. Bumping to 3 minutes will give more headroom.

---
  tests/qtest/meson.build | 1 +
  1 file changed, 1 insertion(+)
tests/qtest/meson.build +0 -0
index c6da428dc5..095c98820e 100644

tests/qtest/meson.build slow_qtests = { · line 11

Quoted 7 lines
tests/qtest/meson.build +1 -0
@@ -5,6 +5,7 @@ slow_qtests = {
    'qom-test' : 900,
    'test-hmp' : 240,
    'pxe-test': 180,
+  'prom-env-test': 180,
  }
 
  qtests_generic = [

I tested your patches, and initially, everything looked good now. But this prom-env-test them reminded me that we run some additional tests in the "SPEED=slow" mode ... I guess we have to take these into consideration, too?

I now did a "configure --enable-debug" build and then ran:

make -j$(nproc) check-qtest-ppc64 SPEED=slow

and indeed, this prom-env-test is now timing out there. Also the device-introspect-test was timing out in SPEED=slow mode. Should we bump the timeout for those, or could this maybe be handled via the TIMEOUT_MULTIPLIER in the final patch?

Thomas

#15 lore raw

Re: [PATCH v2 07/12] qtest: bump prom-env-test timeout to 3 minutes

Quoted 23 lines · Thomas Huth
Quoted 18 lines · Daniel P. Berrangé

The prom-env-test takes about 1 + 1/2 minutes in a --enable-debug build. Bumping to 3 minutes will give more headroom.

---
  tests/qtest/meson.build | 1 +

  1 file changed, 1 insertion(+)
tests/qtest/meson.build +1 -0
index c6da428dc5..095c98820e 100644
@@ -5,6 +5,7 @@ slow_qtests = {
     'qom-test' : 900,
     'test-hmp' : 240,
     'pxe-test': 180,
+  'prom-env-test': 180,
   }
Quoted 1 line

  qtests_generic = [

I tested your patches, and initially, everything looked good now. But this prom-env-test them reminded me that we run some additional tests in the "SPEED=slow" mode ... I guess we have to take these into consideration, too?

Quoted 5 lines

I now did a "configure --enable-debug" build and then ran:

 make -j$(nproc) check-qtest-ppc64 SPEED=slow

and indeed, this prom-env-test is now timing out there. Also the device-introspect-test was timing out in SPEED=slow mode. Should we bump the timeout for those, or could this maybe be handled via the TIMEOUT_MULTIPLIER in the final patch?

I've now done a full check-qtest run with SPEED=slow and --enable-debug, and these tests are timing out:

 14/131 qemu:qtest+qtest-aarch64 / qtest-aarch64/migration-test            TIMEOUT        300.02s   killed by signal 15 SIGTERM
 89/131 qemu:qtest+qtest-s390x / qtest-s390x/device-introspect-test        TIMEOUT         60.01s   killed by signal 15 SIGTERM
103/131 qemu:qtest+qtest-ppc64 / qtest-ppc64/device-introspect-test        TIMEOUT         60.02s   killed by signal 15 SIGTERM
108/131 qemu:qtest+qtest-ppc64 / qtest-ppc64/pxe-test                      TIMEOUT        180.02s   killed by signal 15 SIGTERM
129/131 qemu:qtest+qtest-x86_64 / qtest-x86_64/device-introspect-test      TIMEOUT         60.05s   killed by signal 15 SIGTERM
130/131 qemu:qtest+qtest-aarch64 / qtest-aarch64/device-introspect-test    TIMEOUT         60.01s   killed by signal 15 SIGTERM

Thomas

#16 lore raw

Re: [PATCH v2 07/12] qtest: bump prom-env-test timeout to 3 minutes

In reply to Thomas Huth
Quoted 25 lines · Thomas Huth
Quoted 13 lines · Daniel P. Berrangé

The prom-env-test takes about 1 + 1/2 minutes in a --enable-debug build. Bumping to 3 minutes will give more headroom.

---
  tests/qtest/meson.build | 1 +
  1 file changed, 1 insertion(+)
tests/qtest/meson.build +0 -0
index c6da428dc5..095c98820e 100644

tests/qtest/meson.build slow_qtests = { · line 10

Quoted 6 lines
tests/qtest/meson.build +1 -0
@@ -5,6 +5,7 @@ slow_qtests = {
    'qom-test' : 900,
    'test-hmp' : 240,
    'pxe-test': 180,
+  'prom-env-test': 180,
  }
  qtests_generic = [

I tested your patches, and initially, everything looked good now. But this prom-env-test them reminded me that we run some additional tests in the "SPEED=slow" mode ... I guess we have to take these into consideration, too?

Quoted 8 lines

I now did a "configure --enable-debug" build and then ran:

make -j$(nproc) check-qtest-ppc64 SPEED=slow

and indeed, this prom-env-test is now timing out there. Also the device-introspect-test was timing out in SPEED=slow mode. Should we bump the timeout for those, or could this maybe be handled via the TIMEOUT_MULTIPLIER in the final patch?

My goal was "no regressions" wrt current status for people working on common development hardware. If people are using old, or modern low performance hardware than a timeout multiplier would be expected.

IOW, for SPEED=slow, I think we need to "do the right thing" with the default timeouts.

With regards,
Daniel

Signature · 3 lines
|: https://berrange.com      -o-    https://www.flickr.com/photos/dberrange :|
|: https://libvirt.org         -o-            https://fstop138.berrange.com :|
|: https://entangle-photo.org    -o-    https://www.instagram.com/dberrange :|
#17 lore raw

PATCH v2 8/12 qtest: bump boot-serial-test timeout to 3 minutes

In reply to Daniel P. Berrangé

The boot-serial-test takes about 1 + 1/2 minutes in a --enable-debug build. Bumping to 3 minutes will give more headroom.

---
 tests/qtest/meson.build | 1 +
 1 file changed, 1 insertion(+)
tests/qtest/meson.build +1 -0
index 095c98820e..9e4f5e0ecc 100644
@@ -6,6 +6,7 @@ slow_qtests = {
   'test-hmp' : 240,
   'pxe-test': 180,
   'prom-env-test': 180,
+  'boot-serial-test': 180,
 }
 
 qtests_generic = [
Signature · 1 line
2.41.0
Reviewed-by on v2 8/12
#18 lore raw

Re: [PATCH v2 08/12] qtest: bump boot-serial-test timeout to 3 minutes

Quoted 13 lines · Daniel P. Berrangé

The boot-serial-test takes about 1 + 1/2 minutes in a --enable-debug build. Bumping to 3 minutes will give more headroom.

---
  tests/qtest/meson.build | 1 +
  1 file changed, 1 insertion(+)
tests/qtest/meson.build +0 -0
index 095c98820e..9e4f5e0ecc 100644

tests/qtest/meson.build slow_qtests = { · line 12

Quoted 7 lines
tests/qtest/meson.build +1 -0
@@ -6,6 +6,7 @@ slow_qtests = {
    'test-hmp' : 240,
    'pxe-test': 180,
    'prom-env-test': 180,
+  'boot-serial-test': 180,
  }
 
  qtests_generic = [
#19 lore raw

PATCH v2 9/12 qtest: bump qos-test timeout to 2 minutes

In reply to Daniel P. Berrangé

The qos-test takes just under 1 minute in a --enable-debug build. Bumping to 2 minutes will give more headroom.

---
 tests/qtest/meson.build | 1 +
 1 file changed, 1 insertion(+)
tests/qtest/meson.build +1 -0
index 9e4f5e0ecc..3df9a51b03 100644
@@ -7,6 +7,7 @@ slow_qtests = {
   'pxe-test': 180,
   'prom-env-test': 180,
   'boot-serial-test': 180,
+  'qos-test': 120,
 }
 
 qtests_generic = [
Signature · 1 line
2.41.0
Reviewed-by on v2 9/12
#20 lore raw

Re: [PATCH v2 09/12] qtest: bump qos-test timeout to 2 minutes

Quoted 13 lines · Daniel P. Berrangé

The qos-test takes just under 1 minute in a --enable-debug build. Bumping to 2 minutes will give more headroom.

---
  tests/qtest/meson.build | 1 +
  1 file changed, 1 insertion(+)
tests/qtest/meson.build +0 -0
index 9e4f5e0ecc..3df9a51b03 100644

tests/qtest/meson.build slow_qtests = { · line 13

Quoted 7 lines
tests/qtest/meson.build +1 -0
@@ -7,6 +7,7 @@ slow_qtests = {
    'pxe-test': 180,
    'prom-env-test': 180,
    'boot-serial-test': 180,
+  'qos-test': 120,
  }
 
  qtests_generic = [
#21 lore raw

PATCH v2 10/12 qtest: bump aspeed_smc-test timeout to 4 minutes

In reply to Daniel P. Berrangé

On a reasonably old laptop this test takes 2 minutes 20 seconds with the arm emulator. Raising the timeout to 4 minutes gives greater headroom for slowdown.

---
 tests/qtest/meson.build | 1 +
 1 file changed, 1 insertion(+)
tests/qtest/meson.build +1 -0
index 3df9a51b03..25732a1cdb 100644
@@ -1,4 +1,5 @@
 slow_qtests = {
+  'aspeed_smc-test': 240,
   'bios-tables-test' : 120,
   'migration-test' : 300,
   'npcm7xx_pwm-test': 300,
Signature · 1 line
2.41.0
Reviewed-by on v2 10/12
#22 lore raw

Re: [PATCH v2 10/12] qtest: bump aspeed_smc-test timeout to 4 minutes

Quoted 15 lines · Daniel P. Berrangé

On a reasonably old laptop this test takes 2 minutes 20 seconds with the arm emulator. Raising the timeout to 4 minutes gives greater headroom for slowdown.

---
  tests/qtest/meson.build | 1 +
  1 file changed, 1 insertion(+)
tests/qtest/meson.build +0 -0
index 3df9a51b03..25732a1cdb 100644

tests/qtest/meson.build line 5

Quoted 5 lines
tests/qtest/meson.build +1 -0
@@ -1,4 +1,5 @@
  slow_qtests = {
+  'aspeed_smc-test': 240,
    'bios-tables-test' : 120,
    'migration-test' : 300,
    'npcm7xx_pwm-test': 300,
#23 lore raw

PATCH v2 11/12 qtest: bump bios-table-test timeout to 9 minutes

In reply to Daniel P. Berrangé

This is reliably hitting the current 2 minute timeout in GitLab CI, and for the TCI job, it even hits a 6 minute timeout.

---
 tests/qtest/meson.build | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
tests/qtest/meson.build +1 -1
index 25732a1cdb..1a5e7f0a15 100644
@@ -1,6 +1,6 @@
 slow_qtests = {
   'aspeed_smc-test': 240,
-  'bios-tables-test' : 120,
+  'bios-tables-test' : 540,
   'migration-test' : 300,
   'npcm7xx_pwm-test': 300,
   'qom-test' : 900,
Signature · 1 line
2.41.0
#24 lore raw

PATCH v2 12/12 mtest2make: stop disabling meson test timeouts

In reply to Daniel P. Berrangé

The mtest2make.py script passes the arg '-t 0' to 'meson test' which disables all test timeouts. This is a major source of pain when running in GitLab CI and a test gets stuck. It will stall until GitLab kills the CI job. This leaves us with little easily consumable information about the stalled test. The TAP format doesn't show the test name until it is completed, and TAP output from multiple tests it interleaved. So we have to analyse the log to figure out what tests had un-finished TAP output present and thus infer which test case caused the hang. This is very time consuming and error prone.

By allowing meson to kill stalled tests, we get a direct display of what test program got stuck, which lets us more directly focus in on what specific test case within the test program hung.

The other issue with disabling meson test timeouts by default is that it makes it more likely that maintainers inadvertantly introduce slowdowns. For example the recent-ish change that accidentally made migrate-test take 15-20 minutes instead of around 1 minute.

The main risk of this change is that the individual test timeouts might be too short to allow completion in high load scenarios. Thus, there is likely to be some short term pain where we have to bump the timeouts for certain tests to make them reliable enough. The preceeding few patches raised the timeouts for all failures that were immediately apparent in GitLab CI.

Even with the possible short term instability, this should still be a net win for debuggability of failed CI pipelines over the long term.

---
 scripts/mtest2make.py | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)
scripts/mtest2make.py +2 -1
index 179dd54871..eb01a05ddb 100644
@@ -27,7 +27,8 @@ def names(self, base):
 .speed.slow = $(foreach s,$(sort $(filter-out %-thorough, $1)), --suite $s)
 .speed.thorough = $(foreach s,$(sort $1), --suite $s)
 
-.mtestargs = --no-rebuild -t 0
+TIMEOUT_MULTIPLIER = 1
+.mtestargs = --no-rebuild -t $(TIMEOUT_MULTIPLIER)
 ifneq ($(SPEED), quick)
 .mtestargs += --setup $(SPEED)
 endif
Signature · 1 line
2.41.0
Reviewed-by on v2 12/12
#25 lore raw

Re: [PATCH v2 12/12] mtest2make: stop disabling meson test timeouts

Quoted 10 lines · Daniel P. Berrangé

The mtest2make.py script passes the arg '-t 0' to 'meson test' which disables all test timeouts. This is a major source of pain when running in GitLab CI and a test gets stuck. It will stall until GitLab kills the CI job. This leaves us with little easily consumable information about the stalled test. The TAP format doesn't show the test name until it is completed, and TAP output from multiple tests it interleaved. So we have to analyse the log to figure out what tests had un-finished TAP output present and thus infer which test case caused the hang. This is very time consuming and error prone.

Quoted 6 lines

By allowing meson to kill stalled tests, we get a direct display of what test program got stuck, which lets us more directly focus in on what specific test case within the test program hung.

The other issue with disabling meson test timeouts by default is that it makes it more likely that maintainers inadvertantly introduce slowdowns.

"inadvertently"

Quoted 7 lines

For example the recent-ish change that accidentally made migrate-test take 15-20 minutes instead of around 1 minute.

The main risk of this change is that the individual test timeouts might be too short to allow completion in high load scenarios. Thus, there is likely to be some short term pain where we have to bump the timeouts for certain tests to make them reliable enough.

[*] (see below)

Quoted 1 line

The preceeding few patches

"preceding"

Quoted 16 lines

raised the timeouts for all failures that were immediately apparent in GitLab CI.

Even with the possible short term instability, this should still be a net win for debuggability of failed CI pipelines over the long term.

---
  scripts/mtest2make.py | 3 ++-
  1 file changed, 2 insertions(+), 1 deletion(-)
scripts/mtest2make.py +0 -0
index 179dd54871..eb01a05ddb 100644

scripts/mtest2make.py def names(self, base): · line 30

Quoted 5 lines
scripts/mtest2make.py +1 -1
@@ -27,7 +27,8 @@ def names(self, base):
  .speed.slow = $(foreach s,$(sort $(filter-out %-thorough, $1)), --suite $s)
  .speed.thorough = $(foreach s,$(sort $1), --suite $s)
 
-.mtestargs = --no-rebuild -t 0
+TIMEOUT_MULTIPLIER = 1

Maybe:

MESON_TIMEOUT_MULTIPLIER ?= 1

So one can run:

  $ make check MESON_TIMEOUT_MULTIPLIER=0

to tune a modified test timeout limit in [*]?

Regardless,

Quoted 4 lines
scripts/mtest2make.py +1 -0
+.mtestargs = --no-rebuild -t $(TIMEOUT_MULTIPLIER)
  ifneq ($(SPEED), quick)
  .mtestargs += --setup $(SPEED)
  endif
#26 lore raw
In reply to Daniel P. Berrangé
Quoted 10 lines · Daniel P. Berrangé

Perhaps the most painful of all the GitLab CI failures we see are the enforced job timeouts:

"ERROR: Job failed: execution took longer than 1h15m0s seconds"

   https://gitlab.com/qemu-project/qemu/-/jobs/4387047648

when that hits the CI log shows what has *already* run, but figuring out what was currently running (or rather stuck) is an horrendously difficult.

I had this in my tree but I see there are a number of review comments to take into account. Will there be a v3 and do we want it this late in the cycle?

Signature · 2 lines
Alex Bennée
Virtualisation Tech Lead @ Linaro
#27 lore raw
Quoted 10 lines · Alex Bennée
Quoted 7 lines · Daniel P. Berrangé

Perhaps the most painful of all the GitLab CI failures we see are the enforced job timeouts:

"ERROR: Job failed: execution took longer than 1h15m0s seconds"

https://gitlab.com/qemu-project/qemu/-/jobs/4387047648

Quoted 7 lines
Quoted 3 lines

when that hits the CI log shows what has *already* run, but figuring out what was currently running (or rather stuck) is an horrendously difficult.

I had this in my tree but I see there are a number of review comments to take into account. Will there be a v3 and do we want it this late in the cycle?

I think this could maybe cause some false positives in the CI until we fine-tuned all related timeouts, so no, we don't want to have this in the last release candidates of 8.1. We should commit it early in the 8.2 cycle (hoping that Daniel has some spare minutes to release a v3), so we can iron out the remaining issues there.

Thomas

#28 lore raw
Quoted 19 lines · Thomas Huth
Quoted 10 lines · Alex Bennée
Quoted 7 lines · Daniel P. Berrangé

Perhaps the most painful of all the GitLab CI failures we see are the enforced job timeouts:

"ERROR: Job failed: execution took longer than 1h15m0s seconds"

https://gitlab.com/qemu-project/qemu/-/jobs/4387047648

Quoted 7 lines
Quoted 3 lines

when that hits the CI log shows what has *already* run, but figuring out what was currently running (or rather stuck) is an horrendously difficult.

I had this in my tree but I see there are a number of review comments to take into account. Will there be a v3 and do we want it this late in the cycle?

Quoted 5 lines

I think this could maybe cause some false positives in the CI until we fine-tuned all related timeouts, so no, we don't want to have this in the last release candidates of 8.1. We should commit it early in the 8.2 cycle (hoping that Daniel has some spare minutes to release a v3), so we can iron out the remaining issues there.

Agreed, it is safer to wait until 8.2

I've been away on holiday, but will post a v3 after I catchup

With regards,
Daniel

Signature · 3 lines
|: https://berrange.com      -o-    https://www.flickr.com/photos/dberrange :|
|: https://libvirt.org         -o-            https://fstop138.berrange.com :|
|: https://entangle-photo.org    -o-    https://www.instagram.com/dberrange :|