# Tests that might fail, but where further investigation is not warranted,
# e.g. known bug, fix is difficult and pending, or sometimes timing issues
# on particular platforms.
#
# Fields are
# - <seqno>
# - <whatami-RE>
# - <comments>
#
# If test <seqno> fails, but there is an entry in this file for that
# <seqno> and the <whatami-RE> field (a grep(1) regular expression)
# matches the output from admin/whatami, then the test is considered
# "Triaged" not "Failed".
#
# The fields here are <tab> separated to allow a space in the
# second field in particular.
#
# Any line that does not begin with a <seqno> followed by a <tab> is
# treated as a comment.  So lines that begin with a # are comments, as
# are lines that begin with white space (the latter is useful to wrap long
# <comments> fields over multiple lines).
#
# Because the <whatami-RE> field is used with egrep, one needs to
# (backslash) escape both [ and ( when a literal match is needed ... (none
# of the other special grep -E RE characters appear in the whatami output).
#
# Typical whatami output looks like:
# bozo                6.0.0    x86_64  Ubuntu 20.04 (focal) [kernel=5.4.0 py=3.8.10 se=none]
# <hostname>          <PCP version>
#                              <hardware>
#                                      <operating system>
#                                                           <key software versions>
#

098	vm37	fails in QA run, but later passes with check-flakey
102	vm37	fails in QA run, but later passes with check-flakey
135	bozo-vm	sometimes this VM (VMware on Windows) goes to sleep and
		the pmlogger fetches are long-delayed causing the test
		to fail miserably
280	.*	there is a potential collision with the regular running
		of pmlogger_check and this test finds the lock file in
		place and goes down a rabbit hole ... it is just too hard
		to engineer our way around this
708	i386.*FreeBSD 14\.3	the pmi Python module completely messes
		up the pmUnits struct and fills it with garbage
710	i386.*FreeBSD 14\.3	same as 708
878	.*	on some QA hosts this fails sometimes when we don't see
		the first or second event record ... despite extensive
		triage the problem is not understood
946	.*	Avahi confusion results in no discovery, disabled by default.
1080	Ubuntu 2[02]\.	for pcp-atop built -O2 (not -O0) valgrind reports
			Conditional jump or move depends on uninitialised value
			at 0x11F1D2: deviattask (deviate.c:340)
			but this seems bogus and is not seen elsewhere
1171	openSUSE Leap 15.6	bcc module profile => LLVM ERROR: and no details
1173	Fedora 38	pmdabcc cannot get the ucalls module to work here
1179	openSUSE Leap 15.6	bcc module ustat => bpf: Argument list too long
1210	vm[0-9][0-9]	may collide with pmlogger_check and (correctly)
			detects presence of /etc/pcp/pmlogger/lock
1264	vm26	like in CI, this failure is simply a mystery
1264	vm28	like in CI, this failure is simply a mystery
1339	OpenBSD 7\.	there is some race condition between pmcd and the sample
			PMDA when using inet sockects to communicate ... the
			test passes sometimes and fails other times and hand
			testing with dbpmda cannot reproduce a failure
1475	Fedora 39	pmdabpf cannot get the mountsnoop module to work here
1490	Fedora 39	mem leak appears to be from the pile of steaming stuff
			know as Python, not the PCP labelsets code
1490	Fedora 40	ditto
1595	vm20	valgrind issue goes away when linux PMDA compiled -O0
1574	.*	valgrind-reported issue in libpcp_web pmseries functions
1662	.*	valgrind-reported issue in libpcp_web archive discovery
		handling of sources
1673	openSUSE Leap 15.[56]	bashreadline.so does not work here for the
				bpf PMDA
1701	Fedora 38	Python exception in json/decoder.py
1702	Fedora 38	bpftrace PMDA is broken here, check-flakey passes
1704	Fedora 38	bpftrace PMDA is broken here, check-flakey passes
1705	Fedora 38	bpftrace PMDA is broken here, check-flakey passes
1707	Fedora 38	bpftrace PMDA is broken here, check-flakey passes
			(sometimes)
1709	Fedora 38	statsd PMDA is broken here, check-flakey passes
1740	vm21		acct(2) just seems broken here
1724	Fedora 38	bpftrace PMDA is broken here, check-flakey passes
			(sometimes)
1809	i386.*FreeBSD 14\.3	same as 708
1870	i386.*FreeBSD 14\.3	same as 708
1973	Ubuntu 16\.04	Python here is 3.5.2 and there a bunch of what look like
			dict ordering and/or cacheing sequence differences
			and some arithmetic rounding differences ... not
			worth wasting more time on
1988	Ubuntu 16\.04	Python here is 3.5.2 and there a bunch of what look
			arithmetic rounding differences ... not worth wasting
			more time on
1989	Ubuntu 16\.04	Python here is 3.5.2 and the output is not even
			close to the expected output with different timestamps
			and order of stats being reported ... not worth
			wasting more time on
1990	Ubuntu 16\.04	Python here is 3.5.2 and the output is not even
			close to the expected output with different timestamps
			and order of stats being reported ... not worth
			wasting more time on
1991	Ubuntu 16\.04	Python here is 3.5.2 and there are a couple of
			timestamps that are off by one second ... not worth
			wasting more time on
# special QA environment probing test
#
1526	.*		failure indicates the environment probe is active,
			not a QA failure
