znapzend无法接收新的文件系统流:数据集很忙

Question

有两个测试系统运行最后一个Nexenta4.0.3FP4社区.Master和从站，

mbuffer安装在两个主机上，ssh信任在主机之间是正常的。

当在主复制快照上运行znapzend时，没有发送到从系统，这里我的znapzend znapzend调试日志：

能请你解释一下哪里有问题吗？

./znapzend --debug --runonce=oztssdpool/VMS
[Mon Jun  1 10:39:22 2015] [info] refreshing backup plans...
[Mon Jun  1 10:39:23 2015] [info] found a valid backup plan for oztssdpool/VMS...
[Mon Jun  1 10:39:23 2015] [debug] snapshot worker for oztssdpool/VMS spawned (28160)
[Mon Jun  1 10:39:23 2015] [info] creating recursive snapshot on oztssdpool/VMS
# zfs snapshot -r oztssdpool/VMS@2015-06-01-103923
[Mon Jun  1 10:39:23 2015] [debug] snapshot worker for oztssdpool/VMS done (28160)
[Mon Jun  1 10:39:23 2015] [debug] send/receive worker for oztssdpool/VMS spawned (28163)
[Mon Jun  1 10:39:23 2015] [info] starting work on backupSet oztssdpool/VMS
# zfs list -H -r -o name oztssdpool/VMS
[Mon Jun  1 10:39:23 2015] [debug] sending snapshots from oztssdpool/VMS to root@ozteksas1:oztsas1pool/VMS
# zfs list -H -o name -t snapshot -s creation -d 1 oztssdpool/VMS
# ssh -o Compression=yes -o CompressionLevel=1 -o Cipher=arcfour -o batchMode=yes -o ConnectTimeout=30 root@ozteksas1 zfs list -H -o name -t snapshot -s creation -d 1 oztsas1pool/VMS
# ssh -o Compression=yes -o CompressionLevel=1 -o Cipher=arcfour -o batchMode=yes -o ConnectTimeout=30 root@ozteksas1 zfs list -H -o name -t snapshot -s creation -d 1 oztsas1pool/VMS
[Mon Jun  1 10:39:24 2015] [debug] receive process on ozteksas1 spawned (28169)
# zfs send oztssdpool/VMS@2015-06-01-103923|/opt/csw/bin/mbuffer -q -s 128k -W 60 -m 256M -O 'ozteksas1:31338'
# ssh -o Compression=yes -o CompressionLevel=1 -o Cipher=arcfour -o batchMode=yes -o ConnectTimeout=30 root@ozteksas1 '/opt/csw/bin/mbuffer -q -s 128k -W 60 -m 256M -4 -I 31338|zfs recv -F oztsas1pool/VMS'
***cannot receive new filesystem stream: dataset is busy
[Mon Jun  1 10:39:28 2015] [debug] receive process on ozteksas1 done (28169)
Mojo::Reactor::Poll: Read failed: Event "close" failed:  at /opt/znapzend/bin/../lib/ZnapZend/ZFS.pm line 380.***
# ssh -o Compression=yes -o CompressionLevel=1 -o Cipher=arcfour -o batchMode=yes -o ConnectTimeout=30 root@ozteksas1 zfs list -H -o name -t snapshot -s creation -d 1 oztsas1pool/VMS
[Mon Jun  1 10:39:29 2015] [debug] cleaning up snapshots on root@ozteksas1:oztsas1pool/VMS
# zfs list -H -o name -t snapshot -s creation -d 1 oztssdpool/VMS
[Mon Jun  1 10:39:29 2015] [debug] cleaning up snapshots on oztssdpool/VMS
[Mon Jun  1 10:39:29 2015] [info] done with backupset oztssdpool/VMS in 6 seconds
[Mon Jun  1 10:39:29 2015] [debug] send/receive worker for oztssdpool/VMS done (28163)

Answer 1

我自己找到了答案，当(lun映射)发生此错误时，

在取消目标zvol之后，znapzend从昨天开始工作。