processes running in a podman container are visible on the host
by Matthias Apitz
I've built an own podman image which boots a SuSE Linux, starts a
PostgreSQL dataserver and a SSH daemon to login into. The end of the
Dockerfile says:
...
# Start PostgreSQL server and SSH daemon on container start
#
COPY initSunRise.sh /usr/local/bin/initSunRise.sh
RUN chmod 0755 /usr/local/bin/initSunRise.sh
RUN mkdir /home/sisis/dump ; chown sisis:sisis /home/sisis/dump
COPY testdb.dmp.gz /home/sisis/dump
RUN chown sisis:sisis /home/sisis/dump/testdb.dmp.gz
RUN echo "/etc/init.d/postgres start" >> /usr/local/bin/start.sh
RUN echo "/usr/local/bin/initSunRise.sh" >> /usr/local/bin/start.sh
RUN echo "/usr/sbin/sshd -D" >> /usr/local/bin/start.sh
RUN chmod 0755 /usr/local/bin/start.sh
EXPOSE 22 8076 3045
ENTRYPOINT /usr/local/bin/start.sh
Before it starts the SSH daemon it fires up a shellscript /usr/local/bin/initSunRise.sh
which should fill the PostgreSQL server with a database.
When I now start the image on the server (also a SuSE SLES) with
podman run -p 2022:22 -t suse
Starting PostgreSQL: ok
nohup: appending output to 'nohup.out'
I can SSH into the system on port 2022 and see:
ssh -p 2022 root@srap57dxr1
root(a)srap57dxr1.dev.oclc.org's password:
Last login: Fri Jan 12 08:35:16 2024 from 10.0.2.100
3856d0d6b62d:~ #
3856d0d6b62d:~ # ps ax
PID TTY STAT TIME COMMAND
1 pts/0 Ss+ 0:00 /bin/sh -c /usr/local/bin/start.sh
30 ? S 0:00 /usr/local/sisis-pap/pgsql-15.1/bin/postmaster -D /da
33 ? Ss 0:00 postgres: logger
34 ? Ss 0:00 postgres: checkpointer
35 ? Ss 0:00 postgres: background writer
37 ? Ss 0:01 postgres: walwriter
38 ? Ss 0:00 postgres: autovacuum launcher
39 ? Ss 0:00 postgres: logical replication launcher
45 pts/0 S+ 0:00 sh -c gzip -dc /home/sisis/dump/testdb.dmp.gz | /usr/
46 pts/0 S+ 0:00 sshd: /usr/sbin/sshd -D [listener] 0 of 10-100 startu
47 pts/0 S+ 0:09 gzip -dc /home/sisis/dump/testdb.dmp.gz
48 pts/0 S+ 0:04 /usr/local/sisis-pap/pgsql/bin/psql -Usisis sisis
49 ? Rs 1:28 postgres: sisis sisis [local] COPY
51 ? Rs 0:08 postgres: autovacuum worker sisis
52 ? Rs 0:00 sshd: root@pts/1
54 pts/1 Ss 0:00 -bash
This looks all fine 47 is the process which decompresses the dump and
pipes it to the proc 48 which is a SQL process loading the data into the
PostgreSQL server. This looks all fine to me and works as expected.
But, when I now to a 'ps ax' on the host server where the 'podman run ...'
runs, I see the same proceses with other PID as they would run on the
host and not in the container:
6322 pts/0 Sl+ 0:00 podman run -p 2022:22 -t suse
6337 pts/0 S 0:00 /usr/bin/slirp4netns --disable-host-loopback --mtu=65520 --enable-sandbox --enable-seccomp --enable-ipv6 -c -r 3 -e 4 --netns-type=path /data/guru/containers/tmp/netns/netns-23da4d70-34da-9034-917b-1e56362a5bbb tap0
6344 pts/0 Sl 0:00 rootlessport
6354 pts/0 Sl 0:00 rootlessport-child
6363 ? Ss 0:00 /usr/bin/conmon --api-version 1 -c 3856d0d6b62d01b46c17c160937ea6e9b8239ec64ce2c8e3ce3b4648652b47dc -u 3856d0d6b62d01b46c17c160937ea6e9b8239ec64ce2c8e3ce3b4648652b47dc -r /usr/bin/runc -b /home/apitzm/.local/share/containers/storage/overlay-containers/3856d0d6b62d01b46c17c160937ea6e9b8239ec64ce2c8e3ce3b4648652b47dc/userdata -p /home/apitzm/.local/share/containers/tmp/overlay-containers/3856d0d6b62d01b46c17c160937ea6e9b8239ec64ce2c8e3ce3b4648652b47dc/userdata/pidfile -n great_bouman --exit-dir /run/user/200007/libpod/tmp/exits --full-attach -l journald --log-level warning --syslog.....
...
6428 pts/0 S+ 0:00 sh -c gzip -dc /home/sisis/dump/testdb.dmp.gz | /usr/local/sisis-pap/pgsql/bin/psql -Usisis sisis >> /home/sisis/err 2>&1 ; /usr/local/sisis-pap/pgsql/bin/psql -Usisis sisis < ~sisis/sc/PosUpdSeq.sql >> /home/sisis/err 2>&1
6429 pts/0 S+ 0:00 sshd: /usr/sbin/sshd -D [listener] 0 of 10-100 startups
6430 pts/0 S+ 0:17 gzip -dc /home/sisis/dump/testdb.dmp.gz
6431 pts/0 S+ 0:06 /usr/local/sisis-pap/pgsql/bin/psql -Usisis sisis
6432 ? Rs 2:09 postgres: sisis sisis [local] COPY
See the gzip proc 6430 and the SQL proc 6431.
Is this correct so by intention that the processes which run in a
container are visible from outside?
Forgive me a maybe stupid question, I'm doing UNIX for 35 years and
podman for 35 days :-)
Thanks
matthias
--
Matthias Apitz, ✉ guru(a)unixarea.de, http://www.unixarea.de/ +49-176-38902045
Public GnuPG key: http://www.unixarea.de/key.pub
I am not at war with Russia. Я не воюю с Россией.
Ich bin nicht im Krieg mit Russland.
1 year, 7 months
can podman rootless support qemu-user-static for other arch?
by Larry Cai
Hi:
I have a use case to run an ARM based container inside podman on an x86
platform.
see https://github.com/multiarch/qemu-user-static
It works for me using rootful podman env
```
[root@podman-priv /]# podman run --rm --privileged
multiarch/qemu-user-static --reset -p yes[root@podman-priv /]# podman
run --rm -t arm32v6/alpine uname -marmv7l[root@podman-priv /]# podman
run --rm -t arm64v8/fedora uname -maarch64
```
but when I try this in rootless environment (i am not good at
technical detail in low level)
```
$ podman run --rm --privileged docker.io/multiarch/qemu-user-static
--reset -p yes
mount: permission denied (are you root?)
mount: permission denied (are you root?)
```
so my questions:
1. is it possible to enable this for rootless env? if not, can I use
sudo command to run it once
2. is it done in system level or needs to be done for each user if it
is the common env
Rgs, larry
1 year, 7 months
Is Network=host actually insecure with rootless Podman?
by jklaiho＠iki.fi
I've run into highly mysterious and randomly occurring issues with both container-to-host and container-to-LAN PostgreSQL connections. It makes me suspect something is awry with either Podman or Pasta networking, because running the same service (a Django gunicorn process) bare metal under Supervisor using host-to-host or host-to-LAN PostgreSQL connections doesn't exhibit this issue.
This post is not about that, though. To isolate what part of the system is causing the issue, I need to set up a separate test server where I'll try moving away from Pasta networking into host networking instead to verify that everything still works, and if it does, move our servers facing the connection issues to use host networking.
I've read Erik Sjölund's excellent writeup at https://github.com/eriksjolund/podman-networking-docs and the Medium post it linked to claiming that host networking is insecure, at https://medium.com/nttlabs/dont-use-host-network-namespace-f548aeeef575. It deals with Docker and rootful containers, so immediately I'm thinking that the applicability to rootless Podman is limited or nonexistent.
We use iptables and only expose SSH and nginx to the outside world. Any applications we run in Pasta-networked rootless containers get reverse proxied with nginx. We only run the actual applications in containers; things like PostgreSQL, nginx and Redis are all host services. Pasta has been a boon to us by conveniently exposing ports bidirectionally between container and host and preserving request IPs.
The basic question is: is host networking in rootless Podman any less secure than just running the same services uncontainerized, directly on the host OS, as a non-root user? Even if we exposed a rootless containerized service port directly to the outside world without reverse proxying, I don't see how this would be any riskier with host networking than it would be with bare metal. If the risk profile is the same or very nearly the same as that, I consider that sufficient.
1 year, 7 months
