Strange parallel-make behavior with MKDEBUG=yes
Here's the last few lines leading up to the failure. It's curious (at least to me) that it appears to be installing the library in parallel with building the library! # ./build.sh -T /build/netbsd-local/tools/x86_64/amd64 \ -D /build/netbsd-local/dest/amd64 \ -O /build/netbsd-local/obj/amd64 \ -R /build/netbsd-local/release \ -V RELEASEMACHINEDIR=amd64 -V MKDEBUG=yes \ -m amd64 -j32 -x release --- install-compat --- /build/netbsd-local/tools/x86_64/amd64/bin/x86_64--netbsd-install -N /build/netbsd-local/src/etc -c -r -o root -g wheel -m 444 libisns.so.0.0 /build/netbsd-local/dest/amd64/usr/lib/i386/libisns.so.0.0 --- libisns.so.0.0 --- /build/netbsd-local/tools/x86_64/amd64/bin/x86_64--netbsd-gcc -Wl,-x -shared -Wl,-soname,libisns.so.0 -Wl,--warn-shared-textrel -Wl,-Map=libisns.so.0.map -m32 --sysroot=/build/netbsd-local/dest/amd64 -o libisns.so.0.0 -Wl,-rpath,/usr/lib/i386 -L=/usr/lib/i386 -Wl,--whole-archive libisns_pic.a -Wl,--no-whole-archive -m32 -L/build/netbsd-local/obj/amd64/compat/amd64/i386/lib/libpthread -lpthread --- /build/netbsd-local/dest/amd64/usr/lib/i386/libisns.so.0.0 --- x86_64--netbsd-install: libisns.so.0.0: stat: No such file or directory This has happened to me twice now, the same failure but on different libraries. The failure does not occur with -j 1, and it happens only for MKDEBUG=yes Anyone have any clue? - | Paul Goyette | PGP Key fingerprint: | E-mail addresses: | | Customer Service | FA29 0E3B 35AF E8AE 6651 | paul at whooppee.com| | Network Engineer | 0786 F758 55DE 53BA 7731 | pgoyette at juniper.net | | Kernel Developer | | pgoyette at netbsd.org | -
Re: Strange parallel-make behavior with MKDEBUG=yes
Paul Goyette wrote: x86_64--netbsd-install: libisns.so.0.0: stat: No such file or directory This has happened to me twice now, the same failure but on different libraries. The failure does not occur with -j 1, and it happens only for MKDEBUG=yes Anyone have any clue? Sounds like PR 44046, but I thought that bug had been fixed... -- Andreas Gustafsson, g...@gson.org
Re: Strange parallel-make behavior with MKDEBUG=yes
On Fri, 27 Dec 2013, Andreas Gustafsson wrote: Paul Goyette wrote: x86_64--netbsd-install: libisns.so.0.0: stat: No such file or directory This has happened to me twice now, the same failure but on different libraries. The failure does not occur with -j 1, and it happens only for MKDEBUG=yes Anyone have any clue? Sounds like PR 44046, but I thought that bug had been fixed... Apparently it has resurfaced. Since it seems to be semi-random, perhaps it wasn't really fixed after all? I have the entire build.sh log file available if anyone wants to look deeper... - | Paul Goyette | PGP Key fingerprint: | E-mail addresses: | | Customer Service | FA29 0E3B 35AF E8AE 6651 | paul at whooppee.com| | Network Engineer | 0786 F758 55DE 53BA 7731 | pgoyette at juniper.net | | Kernel Developer | | pgoyette at netbsd.org | -
Re: Strange parallel-make behavior with MKDEBUG=yes
Paul Goyette wrote: Apparently it has resurfaced. Since it seems to be semi-random, perhaps it wasn't really fixed after all? That seems unlikely, as not a single one of the more than 300 parallel MKDEBUG=yes builds I have done since src/lib/Makefile 1.188 was committed has hit the problem. But only a few of those 300+ builds are recent, so it is possible that it was fixed but has recently reappeared. -- Andreas Gustafsson, g...@gson.org
Re: Networking issues with -current?
On Wed, 25 Dec 2013, Manuel Bouyer wrote: On Tue, Dec 24, 2013 at 07:02:47PM -0800, Paul Goyette wrote: [...] Since the machine that hangs is my primary machine, I haven't yet taken the opportunity to start debugging. If anyone has any clues on what to look for, please let me know. the first thing I would look at is netstat -m and vmstat -m, to see if there are mbuf allocation failures. OK, just had another incident. Under normal operations of this machine, I have about 530-540 mbufs in use (as reported by netstat -m). When it hung, I had 899 in use, and a ping to a local neighbor reported no buffer space available. Interestingly, vmstat -m reported zero failures. BTW, what are calls to protocol drain routine? These seem to go up very slowly over time, and there were 18 at the time of the failure. I tried again to ifconfig wm0 down and the process hung. I tried to switch back to another xterm session, and it was unable to re-draw the window. - | Paul Goyette | PGP Key fingerprint: | E-mail addresses: | | Customer Service | FA29 0E3B 35AF E8AE 6651 | paul at whooppee.com| | Network Engineer | 0786 F758 55DE 53BA 7731 | pgoyette at juniper.net | | Kernel Developer | | pgoyette at netbsd.org | -