Description of problem: Tested on aarch64 As soon as bind-libs from bind-9.11.25-2.fc34 are installed on the system, apps like firefox, chatty etc. do not start. Debugging this with strace shows a timeout due to crashing the resolve function() which denies the app start in the end. Version-Release number of selected component (if applicable): bind-9.11.25-2.fc34 Example crash: Dez 02 23:29:58 fedorapine systemd-coredump[1750]: Process 1682 (chatty) of user 1000 dumped core. Stack trace of thread 1682: #0 0x0000007fa82cd630 raise (libc.so.6 + 0x38630) #1 0x0000007fa82b981c abort (libc.so.6 + 0x2481c) #2 0x0000007f7e4c51fc log_assert_failed_realm.constprop.0 (libnss_resolve.so.2 + 0xf1fc) #3 0x0000007f7e4b9044 strv_find (libnss_resolve.so.2 + 0x3044) #4 0x0000007f7e4c2734 _nss_resolve_gethostbyname4_r (libnss_resolve.so.2 + 0xc734) #5 0x0000007fa835b208 gaih_inet.constprop.0 (libc.so.6 + 0xc6208) #6 0x0000007fa835bd44 getaddrinfo (libc.so.6 + 0xc6d44) #7 0x0000007fa6e93a3c get_fqhostname (libsasl2.so.3 + 0xfa3c) #8 0x0000007fa6e94088 sasl_client_new (libsasl2.so.3 + 0x10088) #9 0x0000007fa8abdd58 jabber_auth_start_cyrus (libjabber.so.0 + 0x4fd58) O== Solution to get app back to start was downgrading to 9.11.25-1 2020-12-02T23:38:03+0100 SUBDEBUG Downgrade: bind-license-32:9.11.25-1.fc34.noarch 2020-12-02T23:38:03+0100 SUBDEBUG Downgrade: bind-libs-lite-32:9.11.25-1.fc34.aarch64 2020-12-02T23:38:03+0100 SUBDEBUG Downgrade: bind-libs-32:9.11.25-1.fc34.aarch64 2020-12-02T23:38:03+0100 SUBDEBUG Downgrade: bind-utils-32:9.11.25-1.fc34.aarch64
I doubt it can be related, bind-9.11.25-2 did just change documentation rebuild, no code changed in bind source between those packages. Also, bind component changes dig, host and nslookup in bind-utils package. But it does not change in any way *getaddrinfo* calls or *libnss_resolve*. That would be failure of systemd-resolved. Can you recheck downgrade of bind package actually helps? Is bind running on localhost machine? It seems to me this should be moved to systemd component, I think they did rebase not long ago. Would systemctl stop Can you check command 'getent hosts <anyname>' would not crash and deliver expected results? Can you reproduce crashes with 'resolvectl query <anyname>'? Replace <anyname> with names chatty or crashing firefox is using. can you install debuginfo and try more detailed stack trace? :: dnf debuginfo-install systemd dnf install gdb coredumpctl dump And run commands in debugger: (gdb) bt (gdb) quit Can you attach also 'named-checkconf -p' output? Is nameserver 127.0.0.1 in /etc/resolv.conf?
1) downgrade helped. immediately after downgrade apps started again. No relogin or reboot needed. 2) the system changed in the meantime. We had several updates i.e. a glibc update. I will update bind and recheck it for you. Result after update to latest bind: Firefox works again. Chatty works too. I assume, this update fixed it : 2020-12-06T13:03:27+0100 SUBDEBUG Upgraded: glibc-devel-2.32.9000-17.fc34.aarch64 2020-12-06T13:04:12+0100 SUBDEBUG Upgraded: glibc-langpack-de-2.32.9000-17.fc34.aarch64 2020-12-06T13:04:13+0100 SUBDEBUG Upgraded: glibc-common-2.32.9000-17.fc34.aarch64 2020-12-06T13:04:13+0100 SUBDEBUG Upgraded: glibc-langpack-en-2.32.9000-17.fc34.aarch64 2020-12-06T13:04:14+0100 SUBDEBUG Upgraded: glibc-2.32.9000-17.fc34.aarch64 As it looks, a case of inconsistent package states. 3) it is and was a wifi connection with dhcp, so it set the correct nameserver in /etc/resolv.conf cat /etc/resolv.conf # Generated by NetworkManager search fritz.box nameserver 192.168.0.254 nameserver fd00::e228:6dff:fe26:b918