Re: unkillable dpkg-query processes
Hi, It seems that instead of getting stuck in the kernel where I thought it would, the process gets stuck elsewhere and also tends to loop allocating memory until all memory in the machine is exhausted and the OOM killer starts to try and kill processes left and right. at least it runs with 100% CPU, attaching strace to the pid doesn't give any results strace-ing the whole process doesn't result in more useful output, but the hanging processes were killable when they were running under strace... Cheers, Bernd - To unsubscribe from this list: send the line unsubscribe sparclinux in the body of a message to [EMAIL PROTECTED] More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: unkillable dpkg-query processes
Hi, just got linked to this thread, so here's a bit input form me :) 1) system type A Sun Fire 280R, with two CPU boards, each carrying a TI UltraSparc III (Cheetah), and 2 GB of RAM. If you need more info, just say. (Bernd Zeimetz has previously suggested that the problem is linked to the processor type, the USIII.) It seems to hit USIII machines with 2 CPUs in one tray much more hard than US II, but once a month our Ultra60 (running two US II) has the same issues - it got much better with since 179c85ea53bef807621f335767e41e23f86f01df, though. before the mentioned patch it died a few times per day. Seems it got better on the USIII here, too (we have a v880 here, the large version of Josip's machine, with 2x 2 CPUs), but it still dies way too often, just not useable in the current state. 2) compiler used to build kernel and is it SMP? gcc (GCC) 4.1.2 20061115 (prerelease) (Debian 4.1.1-21) same compiler here. Please note that non-SMP kernels do not boot on those US-III machines at all (at least I didn't find a single one which does). Cheers, Bernd - To unsubscribe from this list: send the line unsubscribe sparclinux in the body of a message to [EMAIL PROTECTED] More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: unkillable dpkg-query processes
From: Bernd Zeimetz [EMAIL PROTECTED] Date: Fri, 26 Oct 2007 14:30:21 +0200 at least it runs with 100% CPU, attaching strace to the pid doesn't give any results strace-ing the whole process doesn't result in more useful output, but the hanging processes were killable when they were running under strace... When it runs with 100% CPU that's what makes me suspect it's spinning in the kernel futex code somewhere or similar. One thing I notice in the debian bug report is a mention of libnss-db So I did some testing here and without libnss-db installed, running dpkg-query does not use futexes at all. But once I install libnss-db and enable it (by running 'make' under /var/lib/misc then editing /etc/nsswitch.conf to make 'db' get searched first) indeed dpkg-query starts using futexes via the libnss-db library. Josip, do you guys have libnss-db or similar in use on the buildd machine? For those who can reproduce it an have something like libnss-db enabled, try disabling it. - To unsubscribe from this list: send the line unsubscribe sparclinux in the body of a message to [EMAIL PROTECTED] More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: unkillable dpkg-query processes
On Fri, Oct 26, 2007 at 03:01:24PM -0700, David Miller wrote: One thing I notice in the debian bug report is a mention of libnss-db So I did some testing here and without libnss-db installed, running dpkg-query does not use futexes at all. But once I install libnss-db and enable it (by running 'make' under /var/lib/misc then editing /etc/nsswitch.conf to make 'db' get searched first) indeed dpkg-query starts using futexes via the libnss-db library. Josip, do you guys have libnss-db or similar in use on the buildd machine? lebrun.d.o doesn't have libnss-db installed, neither outside nor inside the chroot, sorry. Both setups have the default /etc/nsswitch.conf that searches 'db' before 'files' for protocols, services, ethers, rpc, but that's it. BTW, would you benefit from having an account on this machine? -- 2. That which causes joy or happiness. - To unsubscribe from this list: send the line unsubscribe sparclinux in the body of a message to [EMAIL PROTECTED] More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: unkillable dpkg-query processes
On Sat, Oct 27, 2007 at 12:30:56AM +0200, Bernd Zeimetz wrote: Josip, do you guys have libnss-db or similar in use on the buildd machine? They have, that's what Debian's userdir-ldap uses. No, I have to correct you, this machine isn't part of that setup (at least not yet). -- 2. That which causes joy or happiness. - To unsubscribe from this list: send the line unsubscribe sparclinux in the body of a message to [EMAIL PROTECTED] More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: unkillable dpkg-query processes
Josip Rodin wrote: On Sat, Oct 27, 2007 at 12:30:56AM +0200, Bernd Zeimetz wrote: Josip, do you guys have libnss-db or similar in use on the buildd machine? They have, that's what Debian's userdir-ldap uses. No, I have to correct you, this machine isn't part of that setup (at least not yet). Oh ok, I stand corrected - thought it would have it. -- Bernd Zeimetz [EMAIL PROTECTED] http://bzed.de/ - To unsubscribe from this list: send the line unsubscribe sparclinux in the body of a message to [EMAIL PROTECTED] More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: unkillable dpkg-query processes
For those who can reproduce it an have something like libnss-db enabled, try disabling it. - disabled it - running vgdisplay killed the machine (wanted to create a new LV for a chroot)... it's not accessible at all anymore, I think the kernel is a 2.6.23-something here, I'll build a recent one and give it a try again Will take some time as I need to build on USII... -- Bernd Zeimetz [EMAIL PROTECTED] http://bzed.de/ - To unsubscribe from this list: send the line unsubscribe sparclinux in the body of a message to [EMAIL PROTECTED] More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: unkillable dpkg-query processes
Josip, do you guys have libnss-db or similar in use on the buildd machine? They have, that's what Debian's userdir-ldap uses. For those who can reproduce it an have something like libnss-db enabled, try disabling it. Will do in a few minutes. -- Bernd Zeimetz [EMAIL PROTECTED] http://bzed.de/ - To unsubscribe from this list: send the line unsubscribe sparclinux in the body of a message to [EMAIL PROTECTED] More majordomo info at http://vger.kernel.org/majordomo-info.html
