Wouter Wijngaards
2019-Jun-12 06:16 UTC
[nsd-users] NSD 4.2.0 intermittent segfaults @ libssl ?
Hi, On 11/06/2019 22:22, PGNet Dev wrote:> On 6/11/19 11:59 AM, A. Schulze wrote: >> Am 11.06.19 um 19:54 schrieb A. Schulze: >>> ? - I don't use TLS for AXFR >> if that matters: I *do* use TSIG-Keys to secure the AXFR > > Here, all the slave AXFRs are configured NOKEY.So it is not SSL.? Then it could be TCP Fast Open (server-side fast open, from the logs I see this is the AXFR server).? Or perhaps the file descriptor events could be mangled, in the event handling.? This patch may help?? Otherwise, perhaps a stack trace of the issue could be useful to have, or a way to reproduce it. https://github.com/NLnetLabs/nsd/commit/75c7e1a2fab265e504e8aa2f93dff195dc5b8394 Best regards, Wouter> _______________________________________________ > nsd-users mailing list > nsd-users at NLnetLabs.nl > https://open.nlnetlabs.nl/mailman/listinfo/nsd-users
Am 12.06.19 um 08:16 schrieb Wouter Wijngaards:> Hi, > > On 11/06/2019 22:22, PGNet Dev wrote: >> On 6/11/19 11:59 AM, A. Schulze wrote: >>> Am 11.06.19 um 19:54 schrieb A. Schulze: >>>> ? - I don't use TLS for AXFR >>> if that matters: I *do* use TSIG-Keys to secure the AXFR >> >> Here, all the slave AXFRs are configured NOKEY. > > So it is not SSL.? Then it could be TCP Fast Open (server-side fast > open, from the logs I see this is the AXFR server).? Or perhaps the file > descriptor events could be mangled, in the event handling.? This patch > may help?? Otherwise, perhaps a stack trace of the issue could be useful > to have, or a way to reproduce it. > > https://github.com/NLnetLabs/nsd/commit/75c7e1a2fab265e504e8aa2f93dff195dc5b8394I applied similiar changes to 4.2.0 but this make things worse. https://andreasschulze.de/tmp/axfr_segfault_fix.patch Jun 12 17:48:57 dnsserver02 kernel: [2448848.071188] show_signal_msg: 7871 callbacks suppressed Jun 12 17:48:57 dnsserver02 kernel: [2448848.071191] nsd[59588]: segfault at 4 ip 00007f09e0d92ba3 sp 00007ffe460a0978 error 6 in libev.so.4.0.0[7f09e0d89000+d000] Jun 12 17:48:57 dnsserver02 kernel: [2448848.071684] nsd[59589]: segfault at 4 ip 00007f09e0d92ba3 sp 00007ffe460a0978 error 6 in libev.so.4.0.0[7f09e0d89000+d000] Jun 12 17:48:57 dnsserver02 kernel: [2448848.072179] nsd[59590]: segfault at 4 ip 00007f09e0d92ba3 sp 00007ffe460a0978 error 6 in libev.so.4.0.0[7f09e0d89000+d000] Jun 12 17:48:57 dnsserver02 kernel: [2448848.072672] nsd[59591]: segfault at 4 ip 00007f09e0d92ba3 sp 00007ffe460a0978 error 6 in libev.so.4.0.0[7f09e0d89000+d000] Jun 12 17:48:57 dnsserver02 kernel: [2448848.073452] nsd[59592]: segfault at 4 ip 00007f09e0d92ba3 sp 00007ffe460a0978 error 6 in libev.so.4.0.0[7f09e0d89000+d000] Jun 12 17:48:57 dnsserver02 kernel: [2448848.073972] nsd[59593]: segfault at 4 ip 00007f09e0d92ba3 sp 00007ffe460a0978 error 6 in libev.so.4.0.0[7f09e0d89000+d000] Jun 12 17:48:57 dnsserver02 kernel: [2448848.075735] nsd[59594]: segfault at 4 ip 00007f09e0d92ba3 sp 00007ffe460a0978 error 6 in libev.so.4.0.0[7f09e0d89000+d000] Jun 12 17:48:57 dnsserver02 kernel: [2448848.076253] nsd[59595]: segfault at 4 ip 00007f09e0d92ba3 sp 00007ffe460a0978 error 6 in libev.so.4.0.0[7f09e0d89000+d000] Jun 12 17:48:57 dnsserver02 kernel: [2448848.077077] nsd[59596]: segfault at 4 ip 00007f09e0d92ba3 sp 00007ffe460a0978 error 6 in libev.so.4.0.0[7f09e0d89000+d000] Jun 12 17:48:57 dnsserver02 kernel: [2448848.077576] nsd[59597]: segfault at 4 ip 00007f09e0d92ba3 sp 00007ffe460a0978 error 6 in libev.so.4.0.0[7f09e0d89000+d000] Andreas
Am 12.06.19 um 08:16 schrieb Wouter Wijngaards:> Otherwise, perhaps a stack trace of the issue could be useful > to have, or a way to reproduce it.I'm unable to produce core files. - ulimit -c unlimited is set - cat /proc/sys/kernel/core_pattern: /core/corefile /core exist outside and inside chroot and is writeable by nsd user nsd started chroot and without chroot nsd crash but: no corefile :-/ Any more magic required? Andreas