Created attachment 640732 [details] output of emerge --info After upgrading from nautilus to octopus (15.2.2) the mgr-process locks up. strace shows an endless loop like this after startup futex(0x7f35cde700e0, FUTEX_WAKE_PRIVATE, 1) = 0 futex(0x7f35cde700d8, FUTEX_WAIT_BITSET_PRIVATE|FUTEX_CLOCK_REALTIME, 0, {tv_sec=1590056968, tv_nsec=655003000}, FUTEX_BITSET_MATCH_ANY) = -1 ETIMEDOUT (Connection timed out) futex(0x7f35cde700e0, FUTEX_WAKE_PRIVATE, 1) = 0 futex(0x7f35cde700d8, FUTEX_WAIT_BITSET_PRIVATE|FUTEX_CLOCK_REALTIME, 0, {tv_sec=1590056968, tv_nsec=660071000}, FUTEX_BITSET_MATCH_ANY) = -1 ETIMEDOUT (Connection timed out) futex(0x7f35cde700e0, FUTEX_WAKE_PRIVATE, 1) = 0 futex(0x7f35cde700d8, FUTEX_WAIT_BITSET_PRIVATE|FUTEX_CLOCK_REALTIME, 0, {tv_sec=1590056968, tv_nsec=665140000}, FUTEX_BITSET_MATCH_ANY) = -1 ETIMEDOUT (Connection timed out) futex(0x7f35cde700e0, FUTEX_WAKE_PRIVATE, 1) = 0 futex(0x7f35cde700d8, FUTEX_WAIT_BITSET_PRIVATE|FUTEX_CLOCK_REALTIME, 0, {tv_sec=1590056968, tv_nsec=670210000}, FUTEX_BITSET_MATCH_ANY) = -1 ETIMEDOUT (Connection timed out) futex(0x7f35cde700e0, FUTEX_WAKE_PRIVATE, 1) = 0 futex(0x7f35cde700d8, FUTEX_WAIT_BITSET_PRIVATE|FUTEX_CLOCK_REALTIME, 0, {tv_sec=1590056968, tv_nsec=675278000}, FUTEX_BITSET_MATCH_ANY^Cstrace: Process 204438 detached <detached ...> Logfile of mgr only shows: 2020-05-21T09:35:19.490+0200 7f35ccf75f00 0 ceph version 15.2.2 (0c857e985a29d90501a285f242ea9c008df49eb8) octopus (stable), process ceph-mgr, pid 204438 2020-05-21T09:35:19.509+0200 7f35ccf75f00 1 mgr[py] Loading python module 'alerts' 2020-05-21T09:35:19.606+0200 7f35ccf75f00 1 mgr[py] Loading python module 'balancer' 2020-05-21T09:35:19.679+0200 7f35ccf75f00 1 mgr[py] Loading python module 'cephadm' 2020-05-21T09:35:19.878+0200 7f35ccf75f00 1 mgr[py] Loading python module 'crash' 2020-05-21T09:35:19.952+0200 7f35ccf75f00 1 mgr[py] Loading python module 'dashboard' 2020-05-21T09:35:20.805+0200 7f35ccf75f00 1 mgr[py] Loading python module 'devicehealth' 2020-05-21T09:35:20.898+0200 7f35ccf75f00 1 mgr[py] Loading python module 'diskprediction_cloud' 2020-05-21T09:35:21.004+0200 7f35ccf75f00 1 mgr[py] Loading python module 'diskprediction_local' ceph health complains that no mgr is running
I would guess that this is the issue: https://tracker.ceph.com/issues/43447 I guess I can add a diskprediction USE flag that forces <scipy-1.4, and removes diskprediction_local if it's off.
AFAIK scipy was pulled by this update
The bug has been closed via the following commit(s): https://gitweb.gentoo.org/repo/gentoo.git/commit/?id=c8077590d4fd447e07508c7e511e9190920f0ca8 commit c8077590d4fd447e07508c7e511e9190920f0ca8 Author: Patrick McLean <patrick.mclean@sony.com> AuthorDate: 2020-05-28 19:08:53 +0000 Commit: Patrick McLean <chutzpah@gentoo.org> CommitDate: 2020-05-28 19:09:22 +0000 sys-cluster/ceph-15.2.2-r1: Revbump, fix bugs #724508 and #724438 Adds a "diskprediction" USE flag to enable diskprediction_local since it forces an old scipy (forcing off python3_8) (bug #724438) Fixes up the systemd unit and adds a tmpfiles entry (bug #724508) Closes: https://bugs.gentoo.org/724508 Closes: https://bugs.gentoo.org/724438 Copyright: Sony Interactive Entertainment Inc. Package-Manager: Portage-2.3.100, Repoman-2.3.22 Signed-off-by: Patrick McLean <chutzpah@gentoo.org> .../{ceph-15.2.2.ebuild => ceph-15.2.2-r1.ebuild} | 30 +++++++++++++++------- .../ceph/files/ceph-15.2.2-systemd-unit.patch | 12 +++++++++ sys-cluster/ceph/files/ceph-tmpfilesd | 1 + sys-cluster/ceph/metadata.xml | 1 + 4 files changed, 35 insertions(+), 9 deletions(-)
