STRLCPY/criu

tests: fix builds on alpine and centos

Install sudo, create test user with ID 1000, install bash,
fix pidfile creation and pidfile chmod.

v2:
 * use sleep to give the criu daemon some time to start up

v3:
 * Andrei is of course right and sleep is not good solution.
   After adding --status-fd support to criu service, this
   is how we now detect that criu is ready.

v4:
 * This was much more complicated than expected which is related
   to the different versions of the tools on the different travis
   test targets. There seems to be a bug in bash on Ubuntu
    https://lists.gnu.org/archive/html/bug-bash/2017-07/msg00039.html
   which prevents using 'read -n1' on Ubuntu. As a workaround
   the result from CRIU's status FD is now read via python.

   Another problem was discovered on alpine with the loop restore test.
   CRIU says to use setsid even if the process is already using setsid.
   As a workaround, still with setsid, this process is now using
   shell-job true for checkpoint and restore.

Parts of v2 have been committed before. So the changes from this commit
are partially already in another commit.

Signed-off-by: Adrian Reber <[email protected]>
Signed-off-by: Andrei Vagin <[email protected]>

Adrian Reber committed with Andrei Vagin 6 years ago

eba1fba8

1 parent ae76eb34

Revision indexing in progress... (symbol navigation in revisions will be accurate after indexed)

Total 5 files Show one by one

■ ■ ■ ■ ■ ■

scripts/build/Dockerfile.centos

		skipped 39 lines
40	40		ENV CCACHE_DIR=/tmp/.ccache CCACHE_NOCOMPRESS=1 $ENV1=yes
41	41		RUN mv .ccache /tmp && make mrproper && ccache -sz && \
42	42		date && make -j $(nproc) CC="$CC" && date && ccache -s
	43	+
	44	+	# The rpc test cases are running as user #1000, let's add the user
	45	+	RUN adduser -u 1000 test
	46	+
43	47		RUN make -C test/zdtm -j $(nproc)
44	48

■ ■ ■ ■ ■ ■

test/others/rpc/Makefile

		skipped 3 lines
4	4		CFLAGS += -g -Werror -Wall -I.
5	5		LDLIBS += -lprotobuf-c
6	6
	7	+	PYTHON ?= python
	8	+
7	9		run: all
8	10		mkdir -p build
9	11		chmod a+rwx build
10		-	@# need to start the criu daemon here to access the pidfile
11		-	sudo -g '#1000' -u '#1000' ./criu service -v4 -W build -o service.log --address criu_service.socket -d --pidfile pidfile
12		-	# Give the criu daemon some time to start up
13		-	sleep 0.5
	12	+	rm -f build/status
	13	+	sudo -g '#1000' -u '#1000' mkfifo build/status
	14	+	@# Need to start the criu daemon here to access the pidfile.
	15	+	@# The script read.py is used to wait until 'criu service'
	16	+	@# is ready. As 'read -n 1' in some releases has a bug and does
	17	+	@# not read correctly a \0, using python is a workaround.
	18	+	sudo -g '#1000' -u '#1000' -- bash -c "exec 200<>build/status; \
	19	+	./criu service -v4 -W build --address criu_service.socket \
	20	+	-d --pidfile pidfile -o service.log --status-fd 200; \
	21	+	$(PYTHON) read.py build/status"
	22	+	rm -f build/status
14	23		chmod a+rw build/pidfile
15	24		sudo -g '#1000' -u '#1000' ./run.sh
16	25		sudo -g '#1000' -u '#1000' ./version.py
		skipped 19 lines

■ ■ ■ ■ ■ ■

test/others/rpc/read.py

1	+	# This script is used to read a single character from CRIU's status FD.
2	+	# That way we know when the CRIU service is ready. CRIU writes a \0 to
3	+	# the status FD.
4	+	# In theory this could be easily done using 'read -n 1' from bash, but
5	+	# but the bash version on Ubuntu has probably the following bug:
6	+	# https://lists.gnu.org/archive/html/bug-bash/2017-07/msg00039.html
7	+
8	+	import os
9	+	import sys
10	+
11	+
12	+	f=open(sys.argv[1])
13	+	r = f.read(1)
14	+
15	+	if r == '\0':
16	+	sys.exit(0)
17	+
18	+	sys.exit(-1)
19	+

■ ■ ■ ■ ■ ■

test/others/rpc/restore-loop.py

		skipped 18 lines
19	19		req = rpc.criu_req()
20	20		req.type = rpc.RESTORE
21	21		req.opts.images_dir_fd = os.open(args['dir'], os.O_DIRECTORY)
	22	+	# As the dumped process is running with setsid this should not
	23	+	# be necessary. There seems to be a problem for this testcase
	24	+	# in combination with alpine's setsid.
	25	+	# The dump is now done with -j and the restore also.
	26	+	req.opts.shell_job = True
22	27
23	28		# Send request
24	29		s.send(req.SerializeToString())
		skipped 17 lines

■ ■ ■ ■ ■ ■

test/others/rpc/run.sh

		skipped 49 lines
50	50		echo "pid ${P}"
51	51
52	52		title_print "Dump loop.sh"
53		-	${CRIU} dump -v4 -o dump-loop.log -D build/imgs_loop -t ${P}
	53	+	# So theoretically '-j' (--shell-job) should not be necessary, but on alpine
	54	+	# this test fails without it.
	55	+	${CRIU} dump -j -v4 -o dump-loop.log -D build/imgs_loop -t ${P}
54	56
55	57		title_print "Run restore-loop"
56	58		./restore-loop.py build/criu_service.socket build/imgs_loop
		skipped 29 lines