I have a problem which reports being full, but actually isn't:

fs2b:~ # du -sh /media/nss/HRZ
638G    /media/nss/HRZ

fs2b:~ # df -h /media/nss/HRZ
Filesystem      Size  Used Avail Use% Mounted on
HRZ             805G  803G     0 100% /media/nss/HRZ

fs2b:~ # nlvm list pool HRZ
 Name=HRZ
  State=Active Type=NSS32 Size=804.99GB Shared=Yes IsSnap=No ADMediaSupport=No
  Used=804.99GB Free=0KB Segs=1 Volumes=1 Snapshots=0 Move=No
  Created: Tue Aug 14 10:13:31 2007
  Pool segments:
   Index     Start        Next     Size  Partition
     1           0  1688205984  804.99GB sdp1.1
  Volumes on this pool:
   Volume     State Mounted   Quota      Used      Free      ADEnabled
   HRZ       Active   Yes      None   802.04GB       0KB        No

fs2b:~ # nlvm list volume HRZ
 Name=HRZ
  Pool=HRZ State=Active Mounted=Yes Shared=Yes
  Mountpoint=/media/nss/HRZ
  Used=802.04GB Avail=0KB Quota=None Purgeable=0KB
  Attributes=Salvage,Compression,UserSpaceRestrictions,DirectoryQuotas
  ReadAheadBlocks=64 PrimaryNameSpace=LONG
  Objects=983903 Files=897724
  BlockSize=4096 ShredCount=1 AuthModelID=1
  SupportedNameSpaces=DOS,MAC,UNIX,LONG
  CreateTime:  Tue Aug 14 10:15:10 2007
  ArchiveTime: Never

When I deleted a 80 GB file, free space temporarily was shown - but about a minute later it was again zero.

I don't find any big files changed the last 24 hours.

The volume is primary of a DST pair; the secondary does not seem to have a problem.

The server is a node in 2-server-cluster, NCS. A few days ago I upgraded one node from OES 2018.3 to OES 24.2. During the upgrade all volumes were on the other node, OES 2018.3. After the upgrade I cluster migrated two volumes to the upgraded node - first a test volume, which showed no problems, then this HRZ volumes. I don't know if there is a connection with the upgrade.

Any ideas whats wrong here and what to do? I am considering a pool verify, but I am hesitating while the system thinks there is no free space - afaik a verify needs to write?