Bug 2328008 - [CephFS-MDS][system-test] MDS failover and FS Volume remove returns EPERM error stating MDS_TRIM or MDS_CACHE_OVERSIZED but no mds trimming issue on corresponding mds
Summary: [CephFS-MDS][system-test] MDS failover and FS Volume remove returns EPERM err...
Keywords:
Status: CLOSED ERRATA
Alias: None
Product: Red Hat Ceph Storage
Classification: Red Hat Storage
Component: CephFS
Version: 8.0
Hardware: All
OS: Linux
unspecified
high
Target Milestone: ---
: 8.1
Assignee: Rishabh Dave
QA Contact: sumr
Rivka Pollack
URL:
Whiteboard:
Depends On:
Blocks: 2351689
TreeView+ depends on / blocked
 
Reported: 2024-11-22 11:06 UTC by sumr
Modified: 2025-06-26 12:19 UTC (History)
8 users (show)

Fixed In Version: ceph-19.2.1-106.el9cp
Doc Type: Bug Fix
Doc Text:
.MDS and FS IDs are now verified during health warning checks for fail commands Previously, the MDS and FS IDs were not checked when executing the `ceph mds fail` and `ceph fs fail` commands. As a result, these commands would fail with a "permission denied" error for healthy MDS or FS instances when another instance in the cluster exhibited health warnings. With this fix, the system now validates the MDS and FS IDs during the health warning check. This change ensures that the `ceph mds fail` and `ceph fs fail` commands succeed for healthy instances, even if other MDS or FS instances in the cluster have health warnings.
Clone Of:
Environment:
Last Closed: 2025-06-26 12:19:44 UTC
Embargoed:


Attachments (Terms of Use)


Links
System ID Private Priority Status Summary Last Updated
Ceph Project Bug Tracker 69694 0 None None None 2025-02-04 18:03:14 UTC
Ceph Project Bug Tracker 69707 0 None None None 2025-02-04 18:03:14 UTC
Red Hat Issue Tracker RHCEPH-10275 0 None None None 2024-11-22 11:10:44 UTC
Red Hat Product Errata RHSA-2025:9775 0 None None None 2025-06-26 12:19:53 UTC

Description sumr 2024-11-22 11:06:06 UTC
Description of problem:
MDS failover on rank-0 mds returns EPERM error as below, but Ceph Health is OK.

[root@rhel94client1 ~]# ceph health detail
HEALTH_OK
[root@rhel94client1 ~]# ceph mds fail cephfs_1.magna022.xclusw
Error EPERM: MDS has one of two health warnings which could extend recovery: MDS_TRIM or MDS_CACHE_OVERSIZED. MDS failover is not recommended since it might cause unexpected file system unavailability. If you wish to proceed, pass --yes-i-really-mean-it



Version-Release number of selected component (if applicable): 19.2.0-53.el9cp


How reproducible:


Steps to Reproduce:
1. Configure CephFS MDS as 2/2 daemons up, 2 standby, 2 hot standby
2. Upgrade to latest Squid build while IO in-progress
3. After few minutes of upgrade, attempt MDS failover on rank-0 MDS.

Actual results: MDS failover returns EPERM Error as below but Ceph HEALTH is OK.

[root@rhel94client1 ~]# ceph health detail
HEALTH_OK

[root@rhel94client1 ~]# ceph mds fail cephfs_1.magna022.xclusw
Error EPERM: MDS has one of two health warnings which could extend recovery: MDS_TRIM or MDS_CACHE_OVERSIZED. MDS failover is not recommended since it might cause unexpected file system unavailability. If you wish to proceed, pass --yes-i-really-mean-it

[root@rhel94client1 ~]# ceph -s
  cluster:
    id:     38f0f738-95d9-11ef-a651-002590fc2a2e
    health: HEALTH_OK

Expected results: 
If rank-0 MDS had one of two warnings MDS_TRIM or MDS_CACHE_OVERSIZED, it should have been reported in ceph -s or ceph health detail. 
OR
If ceph health is OK, then Ceph mds failover should have been allowed without prompting warning message as above.


Additional info:

[root@rhel94client1 ~]# ceph -s
  cluster:
    id:     38f0f738-95d9-11ef-a651-002590fc2a2e
    health: HEALTH_OK
 
  services:
    mon: 3 daemons, quorum magna021,magna023,magna022 (age 68m)
    mgr: magna022.ilqiwl(active, since 73m), standbys: magna021.uhexkq
    mds: 2/2 daemons up, 2 standby, 2 hot standby
    osd: 21 osds: 21 up (since 30m), 21 in (since 3w)
 
  data:
    volumes: 1/1 healthy
    pools:   4 pools, 561 pgs
    objects: 576.65k objects, 623 GiB
    usage:   1.8 TiB used, 17 TiB / 19 TiB avail
    pgs:     471 active+clean+snaptrim_wait
             49  active+clean
             41  active+clean+snaptrim
 
  io:
    client:   5.2 MiB/s rd, 94 MiB/s wr, 321 op/s rd, 215 op/s wr
 

[root@rhel94client1 ~]# ceph fs status
cephfs_1 - 47 clients
========
RANK      STATE                 MDS                ACTIVITY     DNS    INOS   DIRS   CAPS  
 0        active      cephfs_1.magna022.xclusw  Reqs:  109 /s  42.8k  26.1k  1651   8691   
 1        active      cephfs_1.magna028.egaara  Reqs:   37 /s  33.0k  20.2k  1553   7226   
0-s   standby-replay  cephfs_1.magna024.xsvsbn  Evts:  178 /s  57.3k  22.4k  1946      0   
1-s   standby-replay  cephfs_1.magna027.liffty  Evts:  144 /s  44.4k  17.2k  1555      0   
        POOL            TYPE     USED  AVAIL  
cephfs.cephfs_1.meta  metadata  3838M  5580G  
cephfs.cephfs_1.data    data    1742G  5580G  
      STANDBY MDS         
cephfs_1.magna023.bpteaq  
cephfs_1.magna026.kmtgip  
MDS version: ceph version 19.2.0-53.el9cp (677d8728b1c91c14d54eedf276ac61de636606f8) squid (stable)
[root@rhel94client1 ~]# ceph health detail
HEALTH_OK
[root@rhel94client1 ~]# ceph mds fail cephfs_1.magna022.xclusw
Error EPERM: MDS has one of two health warnings which could extend recovery: MDS_TRIM or MDS_CACHE_OVERSIZED. MDS failover is not recommended since it might cause unexpected file system unavailability. If you wish to proceed, pass --yes-i-really-mean-it
[root@rhel94client1 ~]# ceph orch ps --refresh | grep mds
mds.cephfs_1.magna022.xclusw                magna022                    running (19m)    10m ago  28h     115M        -  19.2.0-53.el9cp  e4177168bc51  6de21f253d2d  
mds.cephfs_1.magna023.bpteaq                magna023                    running (18m)    10m ago  28h    19.1M        -  19.2.0-53.el9cp  e4177168bc51  56c00f330ba8  
mds.cephfs_1.magna024.xsvsbn                magna024                    running (17m)    10m ago  28h    59.4M        -  19.2.0-53.el9cp  e4177168bc51  a01dd90042a1  
mds.cephfs_1.magna026.kmtgip                magna026                    running (17m)    10m ago  28h    15.3M        -  19.2.0-53.el9cp  e4177168bc51  c92e5c4b4e05  
mds.cephfs_1.magna027.liffty                magna027                    running (16m)     9m ago  28h    41.4M        -  19.2.0-53.el9cp  e4177168bc51  ddbd9aeebeba  
mds.cephfs_1.magna028.egaara                magna028                    running (15m)    10m ago  28h    69.4M        -  19.2.0-53.el9cp  e4177168bc51  d5d9c87fc79f  

I will add perf dump and mds debug logs to BZ directory and share the link.

Comment 13 errata-xmlrpc 2025-06-26 12:19:44 UTC
Since the problem described in this bug report should be
resolved in a recent advisory, it has been closed with a
resolution of ERRATA.

For information on the advisory (Important: Red Hat Ceph Storage 8.1 security, bug fix, and enhancement updates), and where to find the updated
files, follow the link below.

If the solution does not work for you, open a new bug report.

https://access.redhat.com/errata/RHSA-2025:9775


Note You need to log in before you can comment on or make changes to this bug.