Fix/ota 831/retry download #1001

patriotyk · 2018-11-12T12:30:17Z

No description provided.

codecov-io · 2018-11-12T13:50:40Z

Codecov Report

Merging #1001 into master will increase coverage by 0.09%.
The diff coverage is 98%.

@@            Coverage Diff             @@
##           master    #1001      +/-   ##
==========================================
+ Coverage   81.95%   82.04%   +0.09%     
==========================================
  Files         188      188              
  Lines       13202    13219      +17     
==========================================
+ Hits        10820    10846      +26     
+ Misses       2382     2373       -9

Impacted Files	Coverage Δ
src/libaktualizr/utilities/utils.h	`90.47% <ø> (ø)`	⬆️
src/libaktualizr/primary/sotauptaneclient.h	`83.33% <ø> (ø)`	⬆️
src/libaktualizr/primary/sotauptaneclient.cc	`81.25% <100%> (+0.28%)`	⬆️
src/libaktualizr/primary/reportqueue.cc	`96% <100%> (ø)`	⬆️
src/libaktualizr/uptane/uptane_network_test.cc	`100% <100%> (ø)`	⬆️
src/libaktualizr/http/httpclient.cc	`91.15% <100%> (+3.49%)`	⬆️
src/libaktualizr/uptane/fetcher.cc	`93.87% <100%> (+1.95%)`	⬆️
src/libaktualizr/uptane/uptane_test.cc	`98.63% <100%> (-0.03%)`	⬇️
src/libaktualizr/utilities/utils.cc	`86% <80%> (-0.09%)`	⬇️
src/libaktualizr/storage/invstorage.cc	`90.3% <0%> (-0.05%)`	⬇️
... and 2 more

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update a4c1f05...17e75b6. Read the comment docs.

pattivacek

I think this looks about right. The Jenkins failure is from the hsm_prov test; it's probably transient.

pattivacek · 2018-11-14T10:55:04Z

src/libaktualizr/uptane/uptane_network_test.cc

+  conf.storage.path = temp_dir.Path();
+  conf.provision.expiry_days = "noerrors";
+  conf.provision.primary_ecu_serial = "noerrors";
+  conf.provision.primary_ecu_hardware_id = "download_failure";


Is there a significance to "download_failure" here? I think I created it with the intention that it would be used by the fake server to trigger the download failures.

pattivacek · 2018-11-14T10:57:09Z

tests/fake_http_server/fake_uptane_server.py

@@ -8,7 +8,24 @@

 class Handler(BaseHTTPRequestHandler):
    def do_GET(self):
-        pass
+        if self.path.startswith("/repo/targets/noerrors"):


Should this really be for "noerrors"? Won't this get triggered for all the network tests?

pattivacek · 2018-11-15T08:08:46Z

I merged my large PR, so this needs to be rebased. I fixed one of the flaky tests, though, so at least that's one less thing to worry about.

patriotyk · 2018-11-15T10:35:43Z

I can't look in to the Jenkins

pattivacek · 2018-11-15T10:42:26Z

I can't look in to the Jenkins

Sorry. It's hsm_prov again. I don't understand why it is failing, but here's the log:

114/116 Test #102: hsm_prov_test .......................................***Failed   64.63 sec
> Running /home/laurent/jenkins-slave/jenkins-home/workspace/aktualizr_PR-1001-7RE67M5RGS372BWYW464TROUSU44ZFUJ23LZBJSSS6WN527WU2YA@3/build-coverage/src/aktualizr_primary/aktualizr --config /tmp/tmpgh_d1_b3/conf.d --loglevel 0 --running-mode once
> Running /home/laurent/jenkins-slave/jenkins-home/workspace/aktualizr_PR-1001-7RE67M5RGS372BWYW464TROUSU44ZFUJ23LZBJSSS6WN527WU2YA@3/build-coverage/src/aktualizr_info/aktualizr-info --config /tmp/tmpgh_d1_b3/conf.d
> Running /home/laurent/jenkins-slave/jenkins-home/workspace/aktualizr_PR-1001-7RE67M5RGS372BWYW464TROUSU44ZFUJ23LZBJSSS6WN527WU2YA@3/build-coverage/src/aktualizr_info/aktualizr-info --config /tmp/tmpgh_d1_b3/conf.d
Device has not yet provisioned (as expected). Running aktualizr-cert-provider.
> Running /home/laurent/jenkins-slave/jenkins-home/workspace/aktualizr_PR-1001-7RE67M5RGS372BWYW464TROUSU44ZFUJ23LZBJSSS6WN527WU2YA@3/build-coverage/src/cert_provider/aktualizr-cert-provider -c **** -l / -r -s -u -g /tmp/tmpgh_d1_b3/conf.d/20-sota_hsm_prov.toml
> Running /home/laurent/jenkins-slave/jenkins-home/workspace/aktualizr_PR-1001-7RE67M5RGS372BWYW464TROUSU44ZFUJ23LZBJSSS6WN527WU2YA@3/scripts/export_to_hsm.sh
> Running pkcs11-tool --module=/usr/lib/softhsm/libsofthsm2.so -O --login --pin 1234
> Running softhsm2-util --show-slots
> Running /home/laurent/jenkins-slave/jenkins-home/workspace/aktualizr_PR-1001-7RE67M5RGS372BWYW464TROUSU44ZFUJ23LZBJSSS6WN527WU2YA@3/build-coverage/src/aktualizr_primary/aktualizr --config /tmp/tmpgh_d1_b3/conf.d --loglevel 0 --running-mode once
> Running /home/laurent/jenkins-slave/jenkins-home/workspace/aktualizr_PR-1001-7RE67M5RGS372BWYW464TROUSU44ZFUJ23LZBJSSS6WN527WU2YA@3/build-coverage/src/aktualizr_info/aktualizr-info --config /tmp/tmpgh_d1_b3/conf.d
> Running /home/laurent/jenkins-slave/jenkins-home/workspace/aktualizr_PR-1001-7RE67M5RGS372BWYW464TROUSU44ZFUJ23LZBJSSS6WN527WU2YA@3/build-coverage/src/aktualizr_info/aktualizr-info --config /tmp/tmpgh_d1_b3/conf.d
> Running /home/laurent/jenkins-slave/jenkins-home/workspace/aktualizr_PR-1001-7RE67M5RGS372BWYW464TROUSU44ZFUJ23LZBJSSS6WN527WU2YA@3/build-coverage/src/aktualizr_info/aktualizr-info --config /tmp/tmpgh_d1_b3/conf.d
> Running /home/laurent/jenkins-slave/jenkins-home/workspace/aktualizr_PR-1001-7RE67M5RGS372BWYW464TROUSU44ZFUJ23LZBJSSS6WN527WU2YA@3/build-coverage/src/aktualizr_info/aktualizr-info --config /tmp/tmpgh_d1_b3/conf.d
> Running /home/laurent/jenkins-slave/jenkins-home/workspace/aktualizr_PR-1001-7RE67M5RGS372BWYW464TROUSU44ZFUJ23LZBJSSS6WN527WU2YA@3/build-coverage/src/aktualizr_info/aktualizr-info --config /tmp/tmpgh_d1_b3/conf.d
> Running /home/laurent/jenkins-slave/jenkins-home/workspace/aktualizr_PR-1001-7RE67M5RGS372BWYW464TROUSU44ZFUJ23LZBJSSS6WN527WU2YA@3/build-coverage/src/aktualizr_info/aktualizr-info --config /tmp/tmpgh_d1_b3/conf.d
> Running /home/laurent/jenkins-slave/jenkins-home/workspace/aktualizr_PR-1001-7RE67M5RGS372BWYW464TROUSU44ZFUJ23LZBJSSS6WN527WU2YA@3/build-coverage/src/aktualizr_info/aktualizr-info --config /tmp/tmpgh_d1_b3/conf.d
> Running /home/laurent/jenkins-slave/jenkins-home/workspace/aktualizr_PR-1001-7RE67M5RGS372BWYW464TROUSU44ZFUJ23LZBJSSS6WN527WU2YA@3/build-coverage/src/aktualizr_info/aktualizr-info --config /tmp/tmpgh_d1_b3/conf.d
Provisioning failed: Reading config: "/tmp/tmpgh_d1_b3/conf.d/20-sota_hsm_prov.toml"
Use existing SQL storage: "/tmp/tmpgh_d1_b3/sql.db"
Storage backend: Sqlite
Device ID: ornate-kuerbiskernbrot-619
Primary ecu serial ID: 1dd24927a4a01d29edea6eb731bb22aef856b5df036df7211cd027738ae4e018
Primary ecu hardware ID: 074b1a25f796
Provisioned on server: no
Fetched metadata: no

I restarted it in case it is just transient, but since it's happened two or three times now, it probably isn't.

pattivacek · 2018-11-16T14:06:53Z

I've narrowed the problem down to the changes to httpclient.cc in last commit about fixing valgrind complaints. Before that change, hsm provisioning works. @patriotyk Why was that change necessary?

It may also need a fix for reboot detection. (FYI @lbonn) At least when I test locally, I have to add something like this to my config:

[bootloader]
reboot_sentinel_dir = "/home/patrick/Code/aktualizr/tests/hsm_test/"

Otherwise it complains with boost::filesystem::status: Permission denied: "/var/run/aktualizr-session/need_reboot".

lbonn · 2018-11-16T15:16:49Z

It looks like it is the same curl bug that @OYTIS initially bumped into (see TODO in httpclient.cc).

It's actually not an unknown bug anymore, I tracked it down to curl/curl#2829.

For current distributions, the workaround is still required but can probably be shortened to:

if (pkcs11_key) {
    curlEasySetoptWrapper(curl_put, CURLOPT_SSLENGINE, "pkcs11");
}

This line should be enough in all cases and I suggest we'd put that in an internal wrapper for cloning curl handles to refactor it.

pattivacek · 2018-11-16T16:06:11Z

It looks like it is the same curl bug that @OYTIS initially bumped into (see TODO in httpclient.cc).

Agreed. I'll update this and try to take care of it since i'm already in the middle of it.

Signed-off-by: Patrick Vacek <[email protected]> Signed-off-by: Serhiy Stetskovych <[email protected]>

Signed-off-by: Serhiy Stetskovych <[email protected]> Signed-off-by: Patrick Vacek <[email protected]>

lbonn · 2018-11-16T16:47:39Z

src/libaktualizr/http/httpclient.cc

+  CURL* curl_post = Utils::curlDupHandleWrapper(curl, pkcs11_key);
+  curlEasySetoptWrapper(curl_post, CURLOPT_URL, url.c_str());
+  curlEasySetoptWrapper(curl_post, CURLOPT_POST, 1);
+  if (pkcs11_key) {


Should be unneeded here (in Utils::curlDupHandleWrapper)

Yep, that was a mistake. Fixed.

Now wrapped in a convenience function to make life easier. Signed-off-by: Patrick Vacek <[email protected]>

pattivacek

I think this is done now. Any further thoughts from anyone before we merge it?

patriotyk · 2018-11-19T09:59:57Z

Looks good for me

patriotyk force-pushed the fix/OTA-831/retry-download branch 3 times, most recently from 1300e67 to 97ccbbb Compare November 12, 2018 13:50

patriotyk force-pushed the fix/OTA-831/retry-download branch from 3ea3457 to b550165 Compare November 12, 2018 17:45

pattivacek reviewed Nov 14, 2018

View reviewed changes

patriotyk force-pushed the fix/OTA-831/retry-download branch from b550165 to ea698fa Compare November 15, 2018 09:20

pattivacek force-pushed the fix/OTA-831/retry-download branch 3 times, most recently from b80af8f to 19f2d6b Compare November 16, 2018 10:41

pattivacek and others added 3 commits November 16, 2018 17:25

Retry downloads if they fail for any reason. WIP.

d56a809

Signed-off-by: Patrick Vacek <[email protected]> Signed-off-by: Serhiy Stetskovych <[email protected]>

Finished retry downloads Patrick's work

989920c

Signed-off-by: Serhiy Stetskovych <[email protected]> Signed-off-by: Patrick Vacek <[email protected]>

Fix Valgrind complain

a312fb2

Signed-off-by: Serhiy Stetskovych <[email protected]> Signed-off-by: Patrick Vacek <[email protected]>

pattivacek force-pushed the fix/OTA-831/retry-download branch from 19f2d6b to 1c1236f Compare November 16, 2018 16:25

lbonn reviewed Nov 16, 2018

View reviewed changes

pattivacek force-pushed the fix/OTA-831/retry-download branch from 1c1236f to 7d01357 Compare November 16, 2018 16:48

Workaround for curl handle duplication with pkcs#11 bug.

17e75b6

Now wrapped in a convenience function to make life easier. Signed-off-by: Patrick Vacek <[email protected]>

pattivacek force-pushed the fix/OTA-831/retry-download branch from 7d01357 to 17e75b6 Compare November 16, 2018 16:50

pattivacek approved these changes Nov 19, 2018

View reviewed changes

pattivacek merged commit 67dca3f into master Nov 19, 2018

pattivacek deleted the fix/OTA-831/retry-download branch November 19, 2018 10:01

pattivacek mentioned this pull request Nov 21, 2018

Aktualizr refuses to update to requested image #906

Closed

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Fix/ota 831/retry download #1001

Fix/ota 831/retry download #1001

patriotyk commented Nov 12, 2018

codecov-io commented Nov 12, 2018 •

edited

Loading

pattivacek left a comment

pattivacek Nov 14, 2018 •

edited

Loading

pattivacek Nov 14, 2018

pattivacek commented Nov 15, 2018

patriotyk commented Nov 15, 2018

pattivacek commented Nov 15, 2018

pattivacek commented Nov 16, 2018 •

edited

Loading

lbonn commented Nov 16, 2018

pattivacek commented Nov 16, 2018

lbonn Nov 16, 2018

pattivacek Nov 16, 2018

pattivacek left a comment

patriotyk commented Nov 19, 2018

Fix/ota 831/retry download #1001

Fix/ota 831/retry download #1001

Conversation

patriotyk commented Nov 12, 2018

codecov-io commented Nov 12, 2018 • edited Loading

Codecov Report

pattivacek left a comment

Choose a reason for hiding this comment

pattivacek Nov 14, 2018 • edited Loading

Choose a reason for hiding this comment

pattivacek Nov 14, 2018

Choose a reason for hiding this comment

pattivacek commented Nov 15, 2018

patriotyk commented Nov 15, 2018

pattivacek commented Nov 15, 2018

pattivacek commented Nov 16, 2018 • edited Loading

lbonn commented Nov 16, 2018

pattivacek commented Nov 16, 2018

lbonn Nov 16, 2018

Choose a reason for hiding this comment

pattivacek Nov 16, 2018

Choose a reason for hiding this comment

pattivacek left a comment

Choose a reason for hiding this comment

patriotyk commented Nov 19, 2018

codecov-io commented Nov 12, 2018 •

edited

Loading

pattivacek Nov 14, 2018 •

edited

Loading

pattivacek commented Nov 16, 2018 •

edited

Loading