Bug#584344: speedy-cgi-perl: FTBFS: tests failed
On 20/08/10 at 19:41 +0300, Niko Tyni wrote: found 584344 2.22-12 thanks On Wed, Aug 18, 2010 at 11:21:13PM +0300, Niko Tyni wrote: On Sun, Aug 15, 2010 at 09:45:42PM +0300, Niko Tyni wrote: On Thu, Jun 03, 2010 at 12:01:36PM +0200, Lucas Nussbaum wrote: Source: speedy-cgi-perl Version: 2.22-11 Severity: serious Tags: squeeze sid User: debian...@lists.debian.org Usertags: qa-ftbfs-20100602 qa-ftbfs Justification: FTBFS on amd64 Test Summary Report --- t/detach.t (Wstat: 0 Tests: 2 Failed: 1) Failed test: 2 My best guess is that it's somehow related to bigger pipe buffers on systems with more memory than mine or something like that, but I can't see the problem in the code no matter how much I glare at it. Close enough, the thing that matters here is the socket write buffer size. I suppose the most correct thing to do here would be to check the size of the socket write buffer with SO_SNDBUF (see socket(7)) and set the data and frontend buffer size accordingly. However, I think I'll take the lazy route at least for squeeze and use my earlier patch that just hardcodes bigger numbers. It turns out that this made the first test in detach.t fail on smaller (?) systems, at least armel and mipsel. I've therefore implemented a better fix that adjusts the test according to the socket write buffer size. I'm attaching the new patch and reopening the bug. Lucas, could you please set me up for access to a test system again so that I could confirm that this works there too? Yes, when would be a good time for you? L. -- To UNSUBSCRIBE, email to debian-bugs-rc-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Bug#584344: speedy-cgi-perl: FTBFS: tests failed
found 584344 2.22-12
thanks
On Wed, Aug 18, 2010 at 11:21:13PM +0300, Niko Tyni wrote:
On Sun, Aug 15, 2010 at 09:45:42PM +0300, Niko Tyni wrote:
On Thu, Jun 03, 2010 at 12:01:36PM +0200, Lucas Nussbaum wrote:
Source: speedy-cgi-perl
Version: 2.22-11
Severity: serious
Tags: squeeze sid
User: debian...@lists.debian.org
Usertags: qa-ftbfs-20100602 qa-ftbfs
Justification: FTBFS on amd64
Test Summary Report
---
t/detach.t (Wstat: 0 Tests: 2 Failed: 1)
Failed test: 2
My best guess is that it's somehow related to bigger pipe buffers on
systems with more memory than mine or something like that, but I can't
see the problem in the code no matter how much I glare at it.
Close enough, the thing that matters here is the socket write buffer size.
I suppose the most correct thing to do here would be to check the size
of the socket write buffer with SO_SNDBUF (see socket(7)) and set the
data and frontend buffer size accordingly. However, I think I'll take
the lazy route at least for squeeze and use my earlier patch that just
hardcodes bigger numbers.
It turns out that this made the first test in detach.t fail on smaller
(?) systems, at least armel and mipsel. I've therefore implemented a
better fix that adjusts the test according to the socket write buffer
size.
I'm attaching the new patch and reopening the bug.
Lucas, could you please set me up for access to a test system again
so that I could confirm that this works there too?
--
Niko Tyni nt...@debian.org
Author: Niko Tyni nt...@debian.org
Description: fix the detach test on systems with a large socket buffer size (#584344)
--- speedy-cgi-perl.orig/speedy/t/detach.t
+++ speedy-cgi-perl/speedy/t/detach.t
@@ -6,9 +6,22 @@
use strict;
use IO::File;
+use Socket;
my $smbuf = 8 * 1024;
-my $lgbuf = 512 * 1024;
+my $lgbuf;
+
+# find out the size of the socket write buffer
+# see http://bugs.debian.org/584344
+my $s;
+if (socket($s, AF_UNIX, SOCK_STREAM, 0)
+(my $packed_size = getsockopt($s, SOL_SOCKET, SO_SNDBUF))) {
+$lgbuf = $smbuf + 2 * unpack(I, $packed_size);
+} else {
+warn(socket() or getsockopt() failed: $!);
+}
+
+$lgbuf = 512 * 1024 if $lgbuf 512 * 1024;
my $scr = 't/scripts/detach';
use vars qw(@open_files @pids %children);
@@ -23,7 +36,7 @@
$| = 1; print ; $| = 0;
my $child;
if (($child = open($fh, -|)) == 0) {
- open(F, $ENV{SPEEDY} -- -B$sz $scr |);
+ open(F, $ENV{SPEEDY} -- -B$sz $scr $lgbuf |);
print scalar F;
close(STDOUT);
sleep 60; # Simulate slow drain of output
--- speedy-cgi-perl.orig/speedy/t/scripts/detach
+++ speedy-cgi-perl/speedy/t/scripts/detach
@@ -1,3 +1,6 @@
$| = 1;
+# the data should not fit in the socket write buffer
+# see http://bugs.debian.org/584344
+my $size = shift || (512 * 1024);
print $$\n;
-print 'x' x (500*1024);
+print 'x' x int(500 / 512 * $size);
Processed: Re: Bug#584344: speedy-cgi-perl: FTBFS: tests failed
Processing commands for cont...@bugs.debian.org:
found 584344 2.22-12
Bug #584344 {Done: Niko Tyni nt...@debian.org} [src:speedy-cgi-perl]
speedy-cgi-perl: FTBFS: test failure on systems with a large socket write
buffer size
Bug Marked as found in versions speedy-cgi-perl/2.22-12; no longer marked as
fixed in versions speedy-cgi-perl/2.22-12 and reopened.
thanks
Stopping processing here.
Please contact me if you need assistance.
--
584344: http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=584344
Debian Bug Tracking System
Contact ow...@bugs.debian.org with problems
--
To UNSUBSCRIBE, email to debian-bugs-rc-requ...@lists.debian.org
with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Bug#584344: speedy-cgi-perl: FTBFS: tests failed
retitle 584344 speedy-cgi-perl: FTBFS: test failure on systems with a large socket write buffer size tag 584344 patch thanks On Sun, Aug 15, 2010 at 09:45:42PM +0300, Niko Tyni wrote: On Thu, Jun 03, 2010 at 12:01:36PM +0200, Lucas Nussbaum wrote: Source: speedy-cgi-perl Version: 2.22-11 Severity: serious Tags: squeeze sid User: debian...@lists.debian.org Usertags: qa-ftbfs-20100602 qa-ftbfs Justification: FTBFS on amd64 Test Summary Report --- t/detach.t (Wstat: 0 Tests: 2 Failed: 1) Failed test: 2 My best guess is that it's somehow related to bigger pipe buffers on systems with more memory than mine or something like that, but I can't see the problem in the code no matter how much I glare at it. Close enough, the thing that matters here is the socket write buffer size. The pipe buffers are always 64 kilobytes on Linux, that's hardcoded in the kernel source. The above test failure can be reproduced by setting the sysctl net.core.wmem_default to 47 or so. (The default value on my system is 124928, and it most probably varies between systems depending on their available memory.) The first test in detach.t is basically checking that the persistent speedy backend can serve new requests after it's all done with the previous one, even if the frontend is still busy buffering the data to the client. The second test makes the buffer inside the frontend very small. When the client reads data in slowly, the backend will block on writes and doesn't get freed to handle a new request. However, when net.core.wmem_default is high enough, the data from the backend to the frontend will be buffered in the kernel and the backend can finish. This makes the second test fail. Increasing the amount of data sent from the backend fixes this when it becomes too much to fit in the kernel socket write buffer. The size of the 'large' buffer in detach.t also needs to be increased because the first test requires that all the data fit in it. I suppose the most correct thing to do here would be to check the size of the socket write buffer with SO_SNDBUF (see socket(7)) and set the data and frontend buffer size accordingly. However, I think I'll take the lazy route at least for squeeze and use my earlier patch that just hardcodes bigger numbers. Will revisit this decision in the future if necessary. -- Niko Tyni nt...@debian.org -- To UNSUBSCRIBE, email to debian-bugs-rc-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Processed: Re: Bug#584344: speedy-cgi-perl: FTBFS: tests failed
Processing commands for cont...@bugs.debian.org: retitle 584344 speedy-cgi-perl: FTBFS: test failure on systems with a large socket write buffer size Bug #584344 [src:speedy-cgi-perl] speedy-cgi-perl: FTBFS: tests failed Changed Bug title to 'speedy-cgi-perl: FTBFS: test failure on systems with a large socket write buffer size' from 'speedy-cgi-perl: FTBFS: tests failed' tag 584344 patch Bug #584344 [src:speedy-cgi-perl] speedy-cgi-perl: FTBFS: test failure on systems with a large socket write buffer size Added tag(s) patch. thanks Stopping processing here. Please contact me if you need assistance. -- 584344: http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=584344 Debian Bug Tracking System Contact ow...@bugs.debian.org with problems -- To UNSUBSCRIBE, email to debian-bugs-rc-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Bug#584344: speedy-cgi-perl: FTBFS: tests failed
On Thu, Aug 05, 2010 at 10:25:24PM +0300, Niko Tyni wrote: On Thu, Jun 03, 2010 at 12:01:36PM +0200, Lucas Nussbaum wrote: Source: speedy-cgi-perl Version: 2.22-11 Severity: serious Tags: squeeze sid User: debian...@lists.debian.org Usertags: qa-ftbfs-20100602 qa-ftbfs Justification: FTBFS on amd64 Test Summary Report --- t/detach.t (Wstat: 0 Tests: 2 Failed: 1) Failed test: 2 However, detach.t has failed every time for Lucas so that one is definitely a deterministic bug somewhere. This keeps happening every time for Lucas, but never for me. My best guess is that it's somehow related to bigger pipe buffers on systems with more memory than mine or something like that, but I can't see the problem in the code no matter how much I glare at it. Possibly just growing the buffers in speedy/t/detach.t would help, but it would be nice to actually understand why. Lucas, would it be possible to give me access to a system where the build fails? -- Niko Tyni nt...@debian.org signature.asc Description: Digital signature
Bug#584344: speedy-cgi-perl: FTBFS: tests failed
On 15/08/10 at 21:45 +0300, Niko Tyni wrote: On Thu, Aug 05, 2010 at 10:25:24PM +0300, Niko Tyni wrote: On Thu, Jun 03, 2010 at 12:01:36PM +0200, Lucas Nussbaum wrote: Source: speedy-cgi-perl Version: 2.22-11 Severity: serious Tags: squeeze sid User: debian...@lists.debian.org Usertags: qa-ftbfs-20100602 qa-ftbfs Justification: FTBFS on amd64 Test Summary Report --- t/detach.t (Wstat: 0 Tests: 2 Failed: 1) Failed test: 2 However, detach.t has failed every time for Lucas so that one is definitely a deterministic bug somewhere. This keeps happening every time for Lucas, but never for me. My best guess is that it's somehow related to bigger pipe buffers on systems with more memory than mine or something like that, but I can't see the problem in the code no matter how much I glare at it. Possibly just growing the buffers in speedy/t/detach.t would help, but it would be nice to actually understand why. Lucas, would it be possible to give me access to a system where the build fails? (credentials sent in a private mail) Lucas signature.asc Description: Digital signature
Bug#584344: speedy-cgi-perl: FTBFS: tests failed
tag 584344 - unreproducible thanks On Sun, Aug 15, 2010 at 09:05:21PM +0200, Lucas Nussbaum wrote: On 15/08/10 at 21:45 +0300, Niko Tyni wrote: My best guess is that it's somehow related to bigger pipe buffers on systems with more memory than mine or something like that, but I can't see the problem in the code no matter how much I glare at it. Possibly just growing the buffers in speedy/t/detach.t would help, but it would be nice to actually understand why. This would seem to be the case. The attached patch makes it work again. I doubt there's anything wrong with the actual code, only with the test assumptions of what makes a buffer 'large'. The system where this is reproducible has 16 GB RAM and 300 GB swap, which is quite a bit more than anywhere else where I tried this. I've now got strace output to compare, and I'll try to gain a better understanding of the problem before uploading a fixed package. (credentials sent in a private mail) Thanks! I'm done for now, will ask again if I need another go. -- Niko Tyni nt...@debian.org diff --git a/speedy/t/detach.t b/speedy/t/detach.t index 3154919..8427347 100644 --- a/speedy/t/detach.t +++ b/speedy/t/detach.t @@ -8,7 +8,7 @@ use strict; use IO::File; my $smbuf = 8 * 1024; -my $lgbuf = 512 * 1024; +my $lgbuf = 51200 * 1024; my $scr = 't/scripts/detach'; use vars qw(@open_files @pids %children); diff --git a/speedy/t/scripts/detach b/speedy/t/scripts/detach index 57d1029..72f35f8 100644 --- a/speedy/t/scripts/detach +++ b/speedy/t/scripts/detach @@ -1,3 +1,3 @@ $| = 1; print $$\n; -print 'x' x (500*1024); +print 'x' x (5*1024);
Processed: Re: Bug#584344: speedy-cgi-perl: FTBFS: tests failed
Processing commands for cont...@bugs.debian.org: tag 584344 - unreproducible Bug #584344 [src:speedy-cgi-perl] speedy-cgi-perl: FTBFS: tests failed Removed tag(s) unreproducible. thanks Stopping processing here. Please contact me if you need assistance. -- 584344: http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=584344 Debian Bug Tracking System Contact ow...@bugs.debian.org with problems -- To UNSUBSCRIBE, email to debian-bugs-rc-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Bug#584344: speedy-cgi-perl: FTBFS: tests failed
On Thu, Jun 03, 2010 at 12:01:36PM +0200, Lucas Nussbaum wrote: Source: speedy-cgi-perl Version: 2.22-11 Severity: serious Tags: squeeze sid User: debian...@lists.debian.org Usertags: qa-ftbfs-20100602 qa-ftbfs Justification: FTBFS on amd64 Test Summary Report --- t/detach.t (Wstat: 0 Tests: 2 Failed: 1) Failed test: 2 t/group.t (Wstat: 0 Tests: 6 Failed: 6) Failed tests: 1-6 I can't reproduce either of these myself, and group.t hasn't failed again in the later attempts whose logs are available at http://people.debian.org/~lucas/logs/2010/ However, detach.t has failed every time for Lucas so that one is definitely a deterministic bug somewhere. Will look into this. -- Niko Tyni nt...@debian.org -- To UNSUBSCRIBE, email to debian-bugs-rc-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Bug#584344: speedy-cgi-perl: FTBFS: tests failed
Source: speedy-cgi-perl Version: 2.22-11 Severity: serious Tags: squeeze sid User: debian...@lists.debian.org Usertags: qa-ftbfs-20100602 qa-ftbfs Justification: FTBFS on amd64 Hi, During a rebuild of all packages in sid, your package failed to build on amd64. Relevant part: make[1]: Entering directory `/build/user-speedy-cgi-perl_2.22-11-amd64-9n4BE6/speedy-cgi-perl-2.22/speedy' PERL_DL_NONLAZY=1 SPEEDY=/build/user-speedy-cgi-perl_2.22-11-amd64-9n4BE6/speedy-cgi-perl-2.22/speedy/speedy SPEEDY_BACKENDPROG=/build/user-speedy-cgi-perl_2.22-11-amd64-9n4BE6/speedy-cgi-perl-2.22/speedy_backend/speedy_backend SPEEDY_MODULE=/build/user-speedy-cgi-perl_2.22-11-amd64-9n4BE6/speedy-cgi-perl-2.22// SPEEDY_TIMEOUT=300 /usr/bin/perl -MExtUtils::Command::MM -e test_harness(0, '../blib/lib', '../blib/arch') t/*.t t/alarm.t . ok t/argv.t .. ok t/basic1.t ok t/basic2.t ok t/be_memleak.t ok t/begin_block.t ... ok t/begin_dup.t . ok t/cgi_reset.t . ok t/chdir.t . ok t/detach.t Failed 1/2 subtests t/eagain.t ok t/exit.t .. ok speedy_backend[11669]: t/scripts/group1: No such file or directory speedy_backend[11671]: t/scripts/group2: No such file or directory speedy_backend[11673]: t/scripts/group1: No such file or directory speedy_backend[11675]: t/scripts/group2: No such file or directory speedy_backend[11677]: t/scripts/group1: No such file or directory speedy_backend[11679]: t/scripts/group2: No such file or directory speedy_backend[11681]: t/scripts/group3: No such file or directory speedy_backend[11683]: t/scripts/group3: No such file or directory speedy_backend[11685]: t/scripts/group3: No such file or directory speedy_backend[11691]: t/scripts/group1: No such file or directory t/group.t . Failed 6/6 subtests t/hold_stdio.t ok t/initial_eof.t ... ok t/killfe.t ok t/maxbackend.t ok t/maxruns.t ... ok t/misc.t .. ok t/mungefds.t .. ok t/register_cleanup.t .. ok t/setopt.t ok t/sh_bang.t ... ok t/shutdown.t .. ok t/signal.t ok t/stdin.t . ok t/stdio.t . ok t/stdio2.t ok t/stdio_close.t ... ok t/thread.t ok t/timeout.t ... ok t/touch.t . ok Failed 2/33 test programs. 7/73 subtests failed. t/wrongfd.t ... ok Test Summary Report --- t/detach.t (Wstat: 0 Tests: 2 Failed: 1) Failed test: 2 t/group.t (Wstat: 0 Tests: 6 Failed: 6) Failed tests: 1-6 Files=33, Tests=73, 130 wallclock secs ( 0.11 usr 0.05 sys + 0.73 cusr 2.27 csys = 3.16 CPU) Result: FAIL make[1]: *** [test_dynamic] Error 255 The full build log is available from: http://people.debian.org/~lucas/logs/2010/06/02/speedy-cgi-perl_2.22-11_lsid64.buildlog A list of current common problems and possible solutions is available at http://wiki.debian.org/qa.debian.org/FTBFS . You're welcome to contribute! About the archive rebuild: The rebuild was done on about 50 AMD64 nodes of the Grid'5000 platform, using a clean chroot. Internet was not accessible from the build systems. -- | Lucas Nussbaum | lu...@lucas-nussbaum.net http://www.lucas-nussbaum.net/ | | jabber: lu...@nussbaum.fr GPG: 1024D/023B3F4F | -- To UNSUBSCRIBE, email to debian-bugs-rc-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
