<HTML><BODY><div class="js-helper js-readmsg-msg">
<style type="text/css"></style>
<div>
<base target="_self" href="https://e.mail.ru/">
<div id="style_15742890930714038672_BODY"><div class="class_1574307660">
Thanks for the answer.<br><br>Sent patch as 'Part of' instead of 'Closes'.<br><br><blockquote style="border-left:1px solid #0857A6;margin:10px;padding:0 0 0 10px;">
Четверг, 21 ноября 2019, 1:11 +03:00 от Vladislav Shpilevoy <<a href="mailto:v.shpilevoy@tarantool.org">v.shpilevoy@tarantool.org</a>>:<br>
<br>
<div id="">
<div class="js-helper_mailru_css_attribute_postfix js-readmsg-msg_mailru_css_attribute_postfix">
<style></style>
<div>
<div id="style_15742879122086265882_BODY_mailru_css_attribute_postfix">On 20/11/2019 02:29, Ilya Kosarev wrote:<br>
> Hi!<br>
> <br>
> Thanks for your review.<br>
> <br>
> Did u run tests on exactly this patchset or on the branch,<br>
> <a href="https://github.com/tarantool/tarantool/tree/i.kosarev/gh-4586-fix-quorum-test" target="_blank" rel=" noopener noreferrer">https://github.com/tarantool/tarantool/tree/i.kosarev/gh-4586-fix-quorum-test</a><br>
> which also contains relay: fix join vclock obtainment in <<a href="https://github.com/tarantool/tarantool/commit/83fa385f2e82911dee1a9189f90b1a94c8455023" target="_blank" rel=" noopener noreferrer">https://github.com/tarantool/tarantool/commit/83fa385f2e82911dee1a9189f90b1a94c8455023</a>><br>
> relay_initial_join <<a href="https://github.com/tarantool/tarantool/commit/83fa385f2e82911dee1a9189f90b1a94c8455023" target="_blank" rel=" noopener noreferrer">https://github.com/tarantool/tarantool/commit/83fa385f2e82911dee1a9189f90b1a94c8455023</a>>?<br>
> It is not yet checked in master (and might not get there, as far as<br>
> Georgy has alternative fix as a part of sync replication), but<br>
> is vital for the stability of the test.<br>
> <br>
> On my machine (Ubuntu 18.04.3 LTS) quorum test works perfectly on<br>
> mentioned branch. I use next bash instruction to run it under 'load':<br>
> l=0 ; while ./test-run.py -j20 `for r in {1..64} ; do echo quorum ; done` 2>/dev/null ; do l=$(($l+1)) ; echo ======== $l ============= ; done<br>
I run it on your branch, and it fails. Even without these<br>
complex bash scripts. I run it 5-10 times, and it fails/crashes.<br>
<br>
> <br>
> Anyway, I guess provided problems are not connected with join_vclock<br>
> patch but are mac os specific, as far as i can't reproduce them locally.<br>
> Guess we have some mac os machines, i will ask for access.<br>
<br>
Maybe they are not related. But at least the wrong test results mean,<br>
that the ticket can't be closed. Because ticket is called:<br>
<br>
test: flaky segfault on replication/quorum test under high load<br>
<br>
And it is still flaky.<br>
<br>
For the crash you can open a new ticket.<br>
<br>
> <br>
> It seems to me that wrong results problem is quite easy to handle.<br>
> I have no idea for now how to handle provided segfault, however, i am<br>
> sure it has nothing to do with the segfault mentioned in the issue, as<br>
> far as it was caused by unsafe iteration of anon replicas. Other wrong<br>
> results problems, mentioned in the ticket, are also handled in the<br>
> patchset.<br>
<br>
The problems, mentioned in the ticket, are just comments. Helpers.<br>
The ticket is about flakiness. Not about several concrete fails. If the<br>
test is still flaky, what is a point of closing this issue and opening<br>
exactly the same again?<br>
<br>
> <br>
> Therefore i propose to close this ticket with the provided patchset<br>
> although there are some other problems. Then i will open new issue with<br>
> error info you provided and start to work on it as soon as i get remote<br>
> access to some mac os machine<br>
</div>
</div>
</div>
</div>
</blockquote>
<br>
<br>-- <br>Ilya Kosarev<br>
</div></div>
<base target="_self" href="https://e.mail.ru/">
</div>
</div></BODY></HTML>