Tests performed
Run IOzone with default configuration.
Conclusion
Umount : kernel BUG at fs/nfs/nfs4state.c:130!
After a sequence of 554 mount/umount, the next call to umount is failling.
The bug is reproductible, and we have have two behaviours:
- Umount is taking 0% CPU, and there is a kernel bug
- Umount is taking 99.9% CPU, and there is no more informations
Line 130 of file fs/nfs/nfs4state.c is a kernel bug on a non empty list. Just before there is a loop while (!list_empty(&clp->cl_unused)) {. I suppose the 100% CPU is seen when we are in the loop, 0% CPU when we are out of the loop (on the kernel bug).
Software configuration
- Server :
- Export options are :
/export
<world>(rw,wdelay,insecure,no_root_squash,no_subtree_check,fsid=0)
- mount options, NFSv4 :
rsize=32768,wsize=32768,timeo=14,intr,noauto
- ./iozone -ace -U /mnt/bench_nfs -f /mnt/bench_nfs/fichier_test
- Linux 2.6.12-rc3-CITI_NFS4_ALL-1
hardware configuration
Client
- 2 processors : Intel(R)
- Xeon(TM) CPU 2.80GHz, cache 512 KB
- Total memory: 2Gb
- Ethernet: 1Gb/s link
- Distribution : modified Fedora Core 2
Server
- 4 processors : Intel(R) Xeon(TM) CPU 2.80GHz, cache 512 KB
- Total memory: 24Gb
- Ethernet: 1Gb/s link
- Distribution : modified Fedora Core 2
