https://blogs.gentoo.org/ago/2020/07/04/gentoo-tinderbox/ Issue: net-nds/openldap-2.4.58-r2 fails tests. Discovered on: sparc (internal ref: sparc_at)
Created attachment 751230 [details] build.log build log and emerge --info
ago has: waiting 7 seconds for syncrepl to receive changes... /var/tmp/portage/net-nds/openldap-2.4.58-r2/work/openldap-2.4.58/tests/scripts/test045-syncreplication-proxied: line 668: 50334 Segmentation fault (core dumped) $SLAPD -f $CONF2 -h $URI2 -d $LVL $TIMING >> $LOG2 2>&1 I have: Waiting 15 seconds for syncrepl to receive changes... /var/tmp/portage/net-nds/openldap-2.4.58-r2/work/openldap-2.4.58/tests/scripts/test019-syncreplication-cascade: line 340: 24949 Segmentation fault $SLAPD -f $CONF1 -h $URI1 -d $LVL $TIMING > $LOG1 2>&1
ago, rolf: I'm going to get the 2.5 & 2.6 series into the tree shortly, after that, can you please test: 2.4.59 2.5.9 2.6.0
/var/tmp/portage/net-nds/openldap-2.4.59-r2/work/openldap-2.4.59/tests/scripts/test019-syncreplication-cascade: line 227: 5321 Segmentation fault $SLAPD -f $CONF1 -h $URI1 -d $LVL $TIMING > $LOG1 2>&1
Sometimes I also see this in 2.4.58-r2: Waiting 15 seconds for syncrepl to receive changes... /var/tmp/portage/net-nds/openldap-2.4.58-r2/work/openldap-2.4.58/tests/scripts/test048-syncrepl-multiproxy: line 458: 11323 Segmentation fault $SLAPD -f $CONF1 -h $URI1 -d $LVL $TIMING >> $LOG1 2>&1 ldapsearch failed at provider (255)! /var/tmp/portage/net-nds/openldap-2.4.58-r2/work/openldap-2.4.58/tests/scripts/test048-syncrepl-multiproxy: line 467: kill: (11323) - No such process >>>>> test048-syncrepl-multiproxy failed for bdb (exit 255)
This does not replicate for me on sparc, even with aggressive CFLAGS. All of the following tests pass for me: =net-nds/openldap-2.4.58-r2 =net-nds/openldap-2.4.59-r2 =net-nds/openldap-2.6.1-r1 Can I get a test to see if this is still an issue?
I see the same errors with 2.4.58-r2 and 2.4.59-r2. With 2.6.1-r1 I get: >>>>> Starting test018-syncreplication-persist for mdb... >>>>> Failed test018-syncreplication-persist for mdb after 42 seconds make[1]: *** [Makefile:321: mdb-yes] Error 255 But I don't see a way to get the error details.
(In reply to Rolf Eike Beer from comment #7) > I see the same errors with 2.4.58-r2 and 2.4.59-r2. With 2.6.1-r1 I get: > > >>>>> Starting test018-syncreplication-persist for mdb... > >>>>> Failed test018-syncreplication-persist for mdb after 42 seconds > make[1]: *** [Makefile:321: mdb-yes] Error 255 > > But I don't see a way to get the error details. That is weird. Can I get an emerge --info on the package so I can see if there's a USE or FLAGS difference?
You can get some more details if you look at (in my case): /var/tmp/portage/net-nds/openldap-2.6.2/work/openldap-OPENLDAP_REL_ENG_2_6_2-abi_x86_64.amd64/tests/testout.*.
I have tested and the failures are spurious, and the backtraces end somewhere in poll() or friends. This all sounds like too short timeouts for a slow or busy machine.
(In reply to Rolf Eike Beer from comment #10) > I have tested and the failures are spurious, and the backtraces end > somewhere in poll() or friends. This all sounds like too short timeouts for > a slow or busy machine. Can you try upping the timeouts and see if it helps? I believe these should respect the environment variables SLEEP0, SLEEP1, SLEEP2, and TIMEOUT depending on the test: https://github.com/openldap/openldap/blob/master/tests/scripts/defines.sh#L72
(In reply to matoro from comment #11) > (In reply to Rolf Eike Beer from comment #10) > > I have tested and the failures are spurious, and the backtraces end > > somewhere in poll() or friends. This all sounds like too short timeouts for > > a slow or busy machine. > > Can you try upping the timeouts and see if it helps? I believe these should > respect the environment variables SLEEP0, SLEEP1, SLEEP2, and TIMEOUT > depending on the test: > https://github.com/openldap/openldap/blob/master/tests/scripts/defines.sh#L72 I've now set this in the ebuild too.