Back to bug 1959354

Who When What Removed Added
Juan Miguel Olmo 2021-05-11 11:07:09 UTC Status NEW ASSIGNED
Assignee jolmomar adking
Veera Raghava Reddy 2021-05-13 14:00:24 UTC Target Release 5.1 5.0
CC vereddy
Sebastian Wagner 2021-05-26 10:26:51 UTC Target Release 5.0 5.1
CC sewagner
Vasishta 2021-05-26 14:42:51 UTC Doc Type If docs needed, set a value Known Issue
Sebastian Wagner 2021-05-31 12:49:25 UTC Priority unspecified medium
Red Hat One Jira (issues.redhat.com) 2021-05-31 12:51:52 UTC Link ID Red Hat Issue Tracker RHCEPH-103
Vasishta 2021-06-24 02:43:57 UTC Blocks 1959686
Ranjini M N 2021-06-24 04:54:57 UTC CC adking, rmandyam
Flags needinfo?(adking)
Adam King 2021-06-24 14:53:08 UTC Doc Text Cause: Trying to upgrade with mds service present and no standby mds daemons

Consequence: upgrade will never complete

Workaround (if any): make sure to have at least one standby mds daemon before upgrading using cephadm. Can check this in output of 'ceph fs status <fs-name>' For example with fs named 'test' and 3 mds daemons, 2 of which are standby I see

[ceph: root@vm-00 /]# ceph fs status test
test - 0 clients
====
RANK STATE MDS ACTIVITY DNS INOS DIRS CAPS
0 creating test.vm-00.gdwokm 10 13 12 0
POOL TYPE USED AVAIL
cephfs.test.meta metadata 0 0
cephfs.test.data data 0 0
STANDBY MDS
test.vm-01.nhhjri
test.vm-02.fielli
MDS version: ceph version 17.0.0-5022-g0b43e5f8 (0b43e5f8edbc6f095abded6d2856228d34c7c2d6) quincy (dev)

If I saw no standby daemons I would need to add another before upgrading.

Result: if the standby mds is present, upgrade will complete successfully, otherwise upgrade will get stuck until another mds daemon is added.
Flags needinfo?(adking)
Ranjini M N 2021-06-29 06:13:40 UTC Docs Contact knortema rmandyam
Doc Text Cause: Trying to upgrade with mds service present and no standby mds daemons

Consequence: upgrade will never complete

Workaround (if any): make sure to have at least one standby mds daemon before upgrading using cephadm. Can check this in output of 'ceph fs status <fs-name>' For example with fs named 'test' and 3 mds daemons, 2 of which are standby I see

[ceph: root@vm-00 /]# ceph fs status test
test - 0 clients
====
RANK STATE MDS ACTIVITY DNS INOS DIRS CAPS
0 creating test.vm-00.gdwokm 10 13 12 0
POOL TYPE USED AVAIL
cephfs.test.meta metadata 0 0
cephfs.test.data data 0 0
STANDBY MDS
test.vm-01.nhhjri
test.vm-02.fielli
MDS version: ceph version 17.0.0-5022-g0b43e5f8 (0b43e5f8edbc6f095abded6d2856228d34c7c2d6) quincy (dev)

If I saw no standby daemons I would need to add another before upgrading.

Result: if the standby mds is present, upgrade will complete successfully, otherwise upgrade will get stuck until another mds daemon is added.
.Upgrade of {storage-product} using Cephadm gets stuck if there are no standby MDS daemons

During an upgrade of a {storage-cluster} with an existing MDS service and with no active standby daemons, the process gets stuck.

To workaround this issue, ensure that you have at least one standby MDS daemon before an upgrade through Cephadm.

Run `ceph fs status _FILE_SYSTEM_NAME_` .

If there are no standby daemons, add MDS daemons and then upgrade the storage cluster. The upgrade works as expected when standby daemons are present.
Flags needinfo?(adking)
Adam King 2021-06-29 12:07:35 UTC Flags needinfo?(adking)
Sebastian Wagner 2021-08-11 12:34:18 UTC Status ASSIGNED POST
Ranjini M N 2021-09-03 04:55:53 UTC Doc Text .Upgrade of {storage-product} using Cephadm gets stuck if there are no standby MDS daemons

During an upgrade of a {storage-cluster} with an existing MDS service and with no active standby daemons, the process gets stuck.

To workaround this issue, ensure that you have at least one standby MDS daemon before an upgrade through Cephadm.

Run `ceph fs status _FILE_SYSTEM_NAME_` .

If there are no standby daemons, add MDS daemons and then upgrade the storage cluster. The upgrade works as expected when standby daemons are present.
.Upgrade of {storage-product} using Cephadm gets stuck if there are no standby MDS daemons

During an upgrade of a {storage-product} with an existing MDS service and with no active standby daemons, the process gets stuck.

To workaround this issue, ensure that you have at least one standby MDS daemon before an upgrade through Cephadm.

Run `ceph fs status _FILE_SYSTEM_NAME_` .

If there are no standby daemons, add MDS daemons and then upgrade the storage cluster. The upgrade works as expected when standby daemons are present.
Ken Dreyer (Red Hat) 2021-09-28 20:42:13 UTC Status POST MODIFIED
Fixed In Version ceph-16.2.6-1.el8cp
errata-xmlrpc 2021-09-28 21:21:46 UTC Status MODIFIED ON_QA
Ameena Suhani S H 2021-10-26 07:16:24 UTC QA Contact vashastr sunnagar
Sunil Kumar Nagaraju 2021-12-21 06:55:48 UTC QA Contact sunnagar mgowri
Aron Gunn 2022-01-18 21:30:09 UTC CC agunn
Blocks 2031073
Red Hat Bugzilla 2022-01-31 23:32:11 UTC CC sewagner
Manasa 2022-02-15 05:58:00 UTC Status ON_QA VERIFIED
Ranjini M N 2022-02-16 07:44:56 UTC Flags needinfo?(adking)
Adam King 2022-02-16 18:44:46 UTC Doc Text .Upgrade of {storage-product} using Cephadm gets stuck if there are no standby MDS daemons

During an upgrade of a {storage-product} with an existing MDS service and with no active standby daemons, the process gets stuck.

To workaround this issue, ensure that you have at least one standby MDS daemon before an upgrade through Cephadm.

Run `ceph fs status _FILE_SYSTEM_NAME_` .

If there are no standby daemons, add MDS daemons and then upgrade the storage cluster. The upgrade works as expected when standby daemons are present.
Formerly, upgrade of RHCS using cephadm would be unable to complete if any filesystem did not have a standby mds daemon.

In RHCS 5.1 upgrade can complete without needing to have a standby mds daemon for each filesystem.
Doc Type Known Issue Bug Fix
Flags needinfo?(adking)
Ranjini M N 2022-02-17 10:39:32 UTC Doc Text Formerly, upgrade of RHCS using cephadm would be unable to complete if any filesystem did not have a standby mds daemon.

In RHCS 5.1 upgrade can complete without needing to have a standby mds daemon for each filesystem.
.The cluster can now be upgraded without the requirement of a standby MDS daemon

Previously, {storage-cluster} upgrade would fail if any filesystem did not have a standby Meta Data Server (MDS) daemon.

With this release, you can upgrade the cluster without the requirement of a standby MDS for each filesystem.
Ranjini M N 2022-03-29 06:46:59 UTC Doc Text .The cluster can now be upgraded without the requirement of a standby MDS daemon

Previously, {storage-cluster} upgrade would fail if any filesystem did not have a standby Meta Data Server (MDS) daemon.

With this release, you can upgrade the cluster without the requirement of a standby MDS for each filesystem.
.The cluster can now be upgraded without the requirement of a standby MDS daemon

Previously, {storage-product} upgrade would fail if any filesystem did not have a standby Meta Data Server (MDS) daemon.

With this release, you can upgrade the cluster without the requirement of a standby MDS for each filesystem.
errata-xmlrpc 2022-04-04 08:01:02 UTC Status VERIFIED RELEASE_PENDING
errata-xmlrpc 2022-04-04 10:20:39 UTC Resolution --- ERRATA
Status RELEASE_PENDING CLOSED
Last Closed 2022-04-04 10:20:39 UTC
errata-xmlrpc 2022-04-04 10:21:04 UTC Link ID Red Hat Product Errata RHSA-2022:1174

Back to bug 1959354