New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
one of the tests in tests/system/cli/test_sync.py occasionally hangs #654
Comments
Please run with If you have a "suspect", please run |
I found both of these tests by running with py.test -v in a loop many times. I'll try and see if I can reproduce it for an individual test with -k. |
Ah sorry I overread that you already identified the tests. I will have a look.
…On 19 July 2017 21:19:30 GMT+02:00, "Jelmer Vernooij" ***@***.***> wrote:
I found both of these tests by running with py.test -v in a loop many
times. I'll try and see if I can reproduce it for an individual test
with -k.
--
You are receiving this because you commented.
Reply to this email directly or view it on GitHub:
#654 (comment)
--
Sent from my Android device with K-9 Mail. Please excuse my brevity.
|
I SIGINTed when it hung:
|
I can't reproduce this in a Debian VM. I assume |
I'll try and get the output from the discover call. On the buildds, it's fairly easy to reproduce this, but on my own machine it's quite hard - I had to run it in a loop and it only hung on try 364. |
Is it easier to reproduce if you install vdirsyncer using Debian's packaging infrastructure? I don't really know what to look for here. |
I managed to reproduce it on my workstation(Debian Sid) from vdirsyncer HEAD. I have no idea what influences how frequently the hang happens. |
I'm closing this since nobody has been able to reliably reproduce the bug. |
@untitaker I just came across this again, and I might be able to consistently reproduce it. Are you still willing to take a look at it? My python knowledge is pretty close. |
Do you know how to reproduce it or do you just have a machine where it constantly happens? I don't want to do remote-debugging. |
I am working with ansible and it was happening more frequently in the VM I was testing. My plan was to prepare a repo with the smallest setup to trigger it. Would that work with you? Btw, funny enough, the error happens quite frequently with my home internet but it improved if I used hotspot from my phone. And the same problem, with another test happened for https://github.com/pimutils/khal. |
Sure, i will figure it out. |
Ok, I will try to put the repo here tomorrow 👍 |
@untitaker sorry for the delay, here it is: https://github.com/cedricpim/archlinux-pacaur I can replicate the issue by doing:
It almost always stops in |
BTW I could reproduce this but have no idea. |
I tried to take a look but python is not my language :( |
It hangs at a thread lock... If I attach to the process with gdb or use strace I see that at least. I stopped when I realized that I didn't have the Python extensions for gdb installed. Did you manage to reproduce this on a debian derivative by any chance? |
I didn't manage but I will try to do it tomorrow and write it back. |
@untitaker just tried it on Debian (Jessie), tests pass but the same test seems to hang for a while (but then resumes). So completely hanging might be related to Archlinux :( |
The original report was on Debian, so I think it's not related to Arch. I think this is a very good opportunity to use http://rr-project.org/. If somebody has the time to record something with it I'd be thankful, otherwise I'll get around to it whenever. |
|
Yes, I was able to reproduce it on 64 bit but only very infrequently.
…On 19 January 2018 21:37:24 GMT+11:00, Markus Unterwaditzer ***@***.***> wrote:
just curious @cedricpim @jelmer, were you able to reproduce this issue
on 64-bit machines? I'm asking because I can smoothly reproduce this
issue in the 32-bit Arch VM @cedricpim provided me, but fail to do so
on the 64-bit Arch host.
--
You are receiving this because you were mentioned.
Reply to this email directly or view it on GitHub:
#654 (comment)
--
Jelmer Vernooij <jelmer@jelmer.uk>
https://www.jelmer.uk/
|
After building Python from source in the VM @cedricpim provided me, I can no longer reproduce the issue. I removed some configuration options for faster build times, see here: https://gist.github.com/untitaker/f1f27e73489b6ea4cf3bb53bf37e34d3 |
Well, with 0.18.0 while packaging it for openSUSE/Tumbleweed I got this:
Complete build log to show all versions of packages and steps taken. Do you think it is yet another appearance of this problem or it is something different? Or did I do something wrong? |
@mcepl Can you run The last test fails since that one needs internet access -- but that's just that one. I'll keep in mind that tests should be able to run offline. |
Could you tell me some exact command to run? When I run just |
(With current master)
Apologies in advance for a somewhat vague and incomplete bug report...
Debian buildds regularly hit hangs when running the vdirsyncer tests in tests/system/cli/test_sync.py. (bug report at https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=836667).
I've had trouble reproducing this issue (whether I use eatmydata seems irrelevant), but after running this test file repeatedly under py.test I managed to hit hangs several times in test_multiple_pairs and test_fetch_only_necessary_params.
No idea (yet) as to the root cause.
The text was updated successfully, but these errors were encountered: