Comments

137 Comments

This update has been unpushed.

@clnetbox: I suppose this got mis-directed here, and you meant that for the cockpit update in FEDORA-2019-33e925fa59

I ran this against the cockpit integration tests, which covers a lot of APIs. Each test checks for new violations, too. I also confirmed that this now works with cockpit-tls. Thanks!

I ran this against the cockpit integration tests, which covers a lot of APIs. Each test checks for new violations, too. I also confirmed that this now works with cockpit-tls. Thanks!

Package installs and works fine.

With this version, we now see regressions with mdadm and pcp:

audit: type=1400 audit(1560408406.661:341): avc:  denied  { read } for  pid=7637 comm="mdadm" path="/var/lib/pcp/pmdas/linux/help.dir" dev="dm-0" ino=27031698 scontext=system_u:system_r:mdadm_t:s0 tcontext=system_u:object_r:pcp_var_lib_t:s0 tclass=file permissive=0
audit: type=1400 audit(1560408406.661:341): avc:  denied  { read } for  pid=7637 comm="mdadm" path="/var/lib/pcp/pmdas/linux/help.pag" dev="dm-0" ino=27031699 scontext=system_u:system_r:mdadm_t:s0 tcontext=system_u:object_r:pcp_var_lib_t:s0 tclass=file permissive=0
audit: type=1400 audit(1560408406.702:342): avc:  denied  { read } for  pid=7639 comm="mdadm" path="/var/lib/pcp/pmdas/linux/help.dir" dev="dm-0" ino=27031698 scontext=system_u:system_r:mdadm_t:s0 tcontext=system_u:object_r:pcp_var_lib_t:s0 tclass=file permissive=0

Aside from this SELinux policy update, the only other package from updates-testing which sounds relevant is

 audit                       x86_64 3.0-0.9.20190507gitf58ec40.fc30
                                                          updates-testing 229 k
 audit-libs                  x86_64 3.0-0.9.20190507gitf58ec40.fc30
                                                          updates-testing 106 k

mdadm and pcp didn't get an update recently, and with selinux-policy from current stable (3.14.3-37.fc30) we don't see these.

So -1 due to this regression.

karma

Install&upgrade and all pages work as expected.

karma

I've run my laptop with this for two days, including several reboots. No noticeable regressions. The cockpit tests also work fine against this version, and they cover quite a lot (in particular the happy eyeballs bits).

Tested with cockpit-podman, together with podman 1.4, works fine.

karma

Tested with cockpit-podman, together with containernetworking-plugins, works fine.

BZ#1710604 Login fails with message "Internal error in login process": incorrect protocol: received invalid length prefix
karma

Install/upgrade and all pages work as expected.

karma

Install, upgrade, and all pages work.

karma

Works fine here as well.

karma

Install/upgrade and all cockpit pages work as expected.

Works great, thank you!

Upgrade starts up fine now. I also tested it with kubernetes and cockpit-kubernetes' tests. Thanks!

This still fails, but differently now. On a current and clean F-29, with etcd-3.2.16-6.fc29.x86_64:

# systemctl status -l etcd
● etcd.service - Etcd Server
   Loaded: loaded (/usr/lib/systemd/system/etcd.service; disabled; vendor preset: disabled)
   Active: active (running) since Tue 2019-04-16 02:18:20 EDT; 35s ago
 Main PID: 7797 (etcd)
    Tasks: 9 (limit: 1147)
   Memory: 35.6M
      CPU: 226ms
   CGroup: /system.slice/etcd.service
           └─7797 /usr/bin/etcd --name=default --data-dir=/var/lib/etcd/default.etcd --listen-client-urls=http://localhost:2379

Apr 16 02:18:20 m1.cockpit.lan etcd[7797]: 8e9e05c52164694d received MsgVoteResp from 8e9e05c52164694d at term 2
Apr 16 02:18:20 m1.cockpit.lan etcd[7797]: 8e9e05c52164694d became leader at term 2
Apr 16 02:18:20 m1.cockpit.lan etcd[7797]: raft.node: 8e9e05c52164694d elected leader 8e9e05c52164694d at term 2
Apr 16 02:18:20 m1.cockpit.lan etcd[7797]: setting up the initial cluster version to 3.2
Apr 16 02:18:20 m1.cockpit.lan etcd[7797]: published {Name:default ClientURLs:[http://localhost:2379]} to cluster cdf818194e3a8c32
Apr 16 02:18:20 m1.cockpit.lan systemd[1]: Started Etcd Server.
Apr 16 02:18:20 m1.cockpit.lan etcd[7797]: ready to serve client requests
Apr 16 02:18:20 m1.cockpit.lan etcd[7797]: serving insecure client requests on 127.0.0.1:2379, this is strongly discouraged!
Apr 16 02:18:20 m1.cockpit.lan etcd[7797]: set the initial cluster version to 3.2
Apr 16 02:18:20 m1.cockpit.lan etcd[7797]: enabled capabilities for version 3.2

Somehow dnf upgrade --enablerepo=updates-testing etcd and dnf upgrade --enablerepo=updates-testing --advisory=FEDORA-2019-049b108f40 say "nothing to do", so it seems this isn't yet published. However, I downloaded the build manually with koji download-build, and it fails again:

# systemctl status -l etcd
● etcd.service - Etcd Server
   Loaded: loaded (/usr/lib/systemd/system/etcd.service; disabled; vendor preset: disabled)
   Active: failed (Result: exit-code) since Tue 2019-04-16 02:25:42 EDT; 2min 33s ago
  Process: 23291 ExecStart=/bin/bash -c GOMAXPROCS=$(nproc) /usr/bin/etcd --name="${ETCD_NAME}" --data-dir="${ETCD_DATA_DIR}" --listen-client-urls="${ETCD_LISTEN_>
 Main PID: 23291 (code=exited, status=1/FAILURE)
      CPU: 16ms

Apr 16 02:25:42 m1.cockpit.lan systemd[1]: Starting Etcd Server...
Apr 16 02:25:42 m1.cockpit.lan etcd[23291]: recognized and used environment variable ETCD_ADVERTISE_CLIENT_URLS=http://localhost:2379
Apr 16 02:25:42 m1.cockpit.lan etcd[23291]: conflicting environment variable "ETCD_LISTEN_CLIENT_URLS" is shadowed by corresponding command-line flag (either unset environment variable or disable flag)
Apr 16 02:25:42 m1.cockpit.lan systemd[1]: etcd.service: Main process exited, code=exited, status=1/FAILURE
Apr 16 02:25:42 m1.cockpit.lan systemd[1]: etcd.service: Failed with result 'exit-code'.
Apr 16 02:25:42 m1.cockpit.lan systemd[1]: Failed to start Etcd Server.

Exactly the same failure as in the previous update:

Apr 15 05:13:18 localhost.localdomain bash[4139]: /usr/bin/etcd: line 1: syntax error near unexpected token `newline'
Apr 15 05:13:18 localhost.localdomain bash[4139]: /usr/bin/etcd: line 1: `!<arch>'

full journal, pull request that exhibits this

Note that in the corresponding fedora-29 update, etcd has a regression. I have not tested this on F30.