Bug#923674: systemd: test-stat-util fails with SIGABRT when /dev/random is bind-mounted from /dev/urandom
> from those references: > "You should consider adding more entropy rather than compromising your > system." > or > "One simple way (which I don't really suggest)" Oh, sure, and maybe also "only do it if you know what you are doing". I'm using sbuild, and the bind-mount is only performed in the chroots being used to build packages, not in the host system. The host systems, additionally, are only used to test package building. I think that's safe enough. Thanks.
Bug#923674: systemd: test-stat-util fails with SIGABRT when /dev/random is bind-mounted from /dev/urandom
Am 09.03.19 um 23:20 schrieb Santiago Vila: > On Sat, Mar 09, 2019 at 11:08:15PM +0100, Michael Biebl wrote: >> Am 09.03.19 um 23:00 schrieb Santiago Vila: >>> The reason I would still consider this as a bug (with whatever severity >>> you like) is that symlinking /dev/random to /dev/urandom is a common >>> workaround for lack of entropy, which in turn has traditionally been a >>> common problem for anybody who tries to build a lot of packages in a row. >> >> Can you point to any reference that this is indeed a common practice? >> I think this is a rather bad and dangerous practice, tbh. > > I don't have a reference for how much common it can be, but it is easy > to find such suggestion as a workaround for the lack of entropy problem. > Some examples: > > https://superuser.com/questions/309840/how-can-i-point-dev-random-to-dev-urandom > https://www.ploxiln.net/urandom_random.html from those references: "You should consider adding more entropy rather than compromising your system." or "One simple way (which I don't really suggest)" -- Why is it that all of the instruments seeking intelligent life in the universe are pointed away from Earth? signature.asc Description: OpenPGP digital signature
Bug#923674: systemd: test-stat-util fails with SIGABRT when /dev/random is bind-mounted from /dev/urandom
I confirm that dropping the /dev/urandom line from /etc/schroot/default/fstab makes the package to build again. Now I fear that crypto packages with random-intensive test suites start to FTBFS again... Thanks.
Bug#923674: systemd: test-stat-util fails with SIGABRT when /dev/random is bind-mounted from /dev/urandom
On Sat, Mar 09, 2019 at 11:08:15PM +0100, Michael Biebl wrote: > Am 09.03.19 um 23:00 schrieb Santiago Vila: > > The reason I would still consider this as a bug (with whatever severity > > you like) is that symlinking /dev/random to /dev/urandom is a common > > workaround for lack of entropy, which in turn has traditionally been a > > common problem for anybody who tries to build a lot of packages in a row. > > Can you point to any reference that this is indeed a common practice? > I think this is a rather bad and dangerous practice, tbh. I don't have a reference for how much common it can be, but it is easy to find such suggestion as a workaround for the lack of entropy problem. Some examples: https://superuser.com/questions/309840/how-can-i-point-dev-random-to-dev-urandom https://www.ploxiln.net/urandom_random.html https://kb.vmware.com/s/article/1036980 Thanks.
Bug#923674: systemd: test-stat-util fails with SIGABRT when /dev/random is bind-mounted from /dev/urandom
Am 09.03.19 um 23:00 schrieb Santiago Vila: > The reason I would still consider this as a bug (with whatever severity > you like) is that symlinking /dev/random to /dev/urandom is a common > workaround for lack of entropy, which in turn has traditionally been a > common problem for anybody who tries to build a lot of packages in a row. Can you point to any reference that this is indeed a common practice? I think this is a rather bad and dangerous practice, tbh. -- Why is it that all of the instruments seeking intelligent life in the universe are pointed away from Earth? signature.asc Description: OpenPGP digital signature
Bug#923674: systemd: test-stat-util fails with SIGABRT when /dev/random is bind-mounted from /dev/urandom
retitle 923674 systemd: test-stat-util fails with SIGABRT when /dev/random is bind-mounted from /dev/urandom severity 923674 normal thanks Sorry for this false positive. I think this retitle reflects much better the issue. The reason I would still consider this as a bug (with whatever severity you like) is that symlinking /dev/random to /dev/urandom is a common workaround for lack of entropy, which in turn has traditionally been a common problem for anybody who tries to build a lot of packages in a row. BTW: After building 28500 source packages in buster, I only found another package which failed for this very reason: https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=907191 (Still trying the build in my autobuilders with the fixed fstab but I expect it will work this time). Thanks a lot.
