[bug#27855] [PATCH] gnu: Add rsync service.

[Oleg Pykhalov]

Hello Christopher,

Christopher Baines <address@hidden> writes:

…
>> >> On Thu, 10 Aug 2017 08:18:20 +0100
>> >> Christopher Baines <address@hidden> wrote:
>> >>  
>> >>> I've attached the beginnings of a system test for this. I haven't
>> >>> got as far as actually testing the service directly yet, but I'm
>> >>> going to try testing it, both when running as root, and when not
>> >>> running as root.
>> >>> 
>> >>> A VM start script can be created from this rsync.scm file, by
>> >>> running: guix system vm gnu/tests/rsync.scm
>> >>> 
>> >>> Also, the test itself can be run by doing:
>> >>>   make check-system TESTS=rsync  
>> >>
>> >> I've now got around to actually testing the service. I've attached
>> >> a patch which creates two system tests. One which tests the
>> >> service with the default port, and one with a higher port (I
>> >> randomly picked 2000).
>> >>
>> >> The default port test currently passes, but the higher port one
>> >> currently fails as the rsync service can't create the PID file.
>> >>
>> >> One option to fix this would be to create a /var/run/rsync
>> >> directory which is owned by the rsyncd user, and then create the
>> >> PID file as /var/run/rsync/pid.  
>> >
>> > I'll check it later, if you'll not be faster than me.  :-)
>> >  
>> >> Just let me know if you have any questions?  
>> > …
>> >
>> > I successfully passed you previous patch test, but the new patch one
>> > with two “ports” fails.  
>> 
>> Trying to understand what is going on.  Working on your latest test
>> patch.
>> 
>> This is what I currently have.  It's not differ from your patch except
>> commented test below.  guix/gnu/tests/rsync.scm
>> https://paste.pound-python.org/raw/ZyfPIvOWAbPd7blPRXf7/
>> 
>> I fully commented
>> (test-equal "Test file correctly received from share" …).
>> 
>> (test-assert "Test file copied to share"
>>   …
>>   (zero? ; ERROR HERE
>>    (system* "rsync" "/tmp/input" "localhost::files/")))
>> 
>> Bottom of test log:
>> --8<---------------cut here---------------start------------->8---
>> rsync: mkstemp ".input.Y6OhRW" (in files) failed: Permission denied
>> (13) rsync error: some files/attrs were not transferred (see previous
>> errors) (code 23) at main.c(1178) [sender=3.1.2] QEMU runs as PID 5
>> connected to QEMU's monitor
>> read QEMU monitor prompt
>> connected to guest REPL
>> %%%% Starting test rsync  (Writing full log to "rsync.log")
>> marionette is ready
>> /gnu/store/bm3kvdwh8q3spvhxbmpmcjwz3fng8z5f-rsync-test-builder:1:
>> FAIL Test file copied to share # of expected passes      2
>> # of unexpected failures  1
>> note: keeping build directory `/tmp/guix-build-rsync-test.drv-9'
>> builder for
>> `/gnu/store/1hv9c616m9xkgi549hgbn1y2vr8mnzfr-rsync-test.drv' failed
>> with exit code 1 @
>> build-failed /gnu/store/1hv9c616m9xkgi549hgbn1y2vr8mnzfr-rsync-test.drv
>> - 1 builder for
>> `/gnu/store/1hv9c616m9xkgi549hgbn1y2vr8mnzfr-rsync-test.drv' failed
>> with exit code 1 TOTAL: 1
>> FAIL: /gnu/store/s0rwmr3iha48jbw2b9xkzcqqsamdl21q-rsync-test make:
>> *** [Makefile:5217: check-system] Error 1 --8<---------------cut
>> here---------------end--------------->8---
>
> With the change to add the uid to the rsync configuration file, the
> rsync-with-default-port passes for me. What test are you running here?

It was “rsync-with-default-port”.  With “uid” passed successfully all
tests in “rsync-with-default-port”.

Test 2000 port
==============

The “test-rsync-with-port-2000” fails, because we sync with default
rsync protocol port in this test but we need 2000.

--8<---------------cut here---------------start------------->8---
(test-assert "Test file copied to share"
  …
  (system* "rsync" "/tmp/input" "localhost::files/"))
--8<---------------cut here---------------end--------------->8---

Why I thought about this:
--8<---------------cut here---------------start------------->8---
rsync: failed to connect to localhost (127.0.0.1): Connection refused (111)
rsync error: error in socket IO (code 10) at clientserver.c(125) [sender=3.1.2]
rsync: failed to connect to localhost (127.0.0.1): Connection refused (111)
rsync error: error in socket IO (code 10) at clientserver.c(125) 
[Receiver=3.1.2]
--8<---------------cut here---------------end--------------->8---

Next lines:
--8<---------------cut here---------------start------------->8---
ice-9/eval.scm:387:11: In procedure eval:
ice-9/eval.scm:387:11: In procedure open-file: No such file or directory: 
"/tmp/output"
QEMU runs as PID 5
connected to QEMU's monitor
read QEMU monitor prompt
connected to guest REPL
%%%% Starting test rsync  (Writing full log to "rsync.log")
marionette is ready
/gnu/store/ahmg6h2fnap2znwhl6b4jrjgp830h4dv-rsync-test-builder:1: FAIL Test 
file copied to share
/gnu/store/ahmg6h2fnap2znwhl6b4jrjgp830h4dv-rsync-test-builder:1: FAIL Test 
file correctly received from share
# of expected passes      2
# of unexpected failures  2
note: keeping build directory `/tmp/guix-build-rsync-test.drv-13'
builder for `/gnu/store/9fakmd8dqzdw1ybff6gb00qfzhrwqdis-rsync-test.drv' failed 
with exit code 1
@ build-failed /gnu/store/9fakmd8dqzdw1ybff6gb00qfzhrwqdis-rsync-test.drv - 1 
builder for `/gnu/store/9fakmd8dqzdw1ybff6gb00qfzhrwqdis-rsync-test.drv' failed 
with exit code 1
TOTAL: 1
FAIL: /gnu/store/rmb3hywspr34fwp5xinphaml7c4xjpvk-rsync-test
make: *** [Makefile:5217: check-system] Error 1
--8<---------------cut here---------------end--------------->8---

I only fixed PID error in current moment.

gnu/tests/rsync.scm
https://paste.pound-python.org/raw/D4ccInuJG0g0899GCdy2/

gnu/services/rsync.scm
https://paste.pound-python.org/raw/URDLeIV3XFZ29SjcZYc6/

>> Also, I have (is related?) error on top of test log.
>> --8<---------------cut here---------------start------------->8---
>> address@hidden ~/src/guix$ ./pre-inst-env make check-system
>> TESTS=rsync-with-default-port Compiling Scheme modules...
>> warning: failed to load '(gnu tests admin)':
>> ERROR: In procedure allocate-struct: Wrong type argument in position
>> 2: 8 --8<---------------cut here---------------end--------------->8---
>
> I think I may have seen this error too, no idea what it relates to
> though.

Should we report this as a bug?