Broken mbuffer pipe and hung task when wrong ZFS dst volume configured
wdoekes opened this issue · comments
2017-01-17 17:10:49,985: INFO: Begin copy of 'virtio0' (mc11-5-local-ssd:vm-109-disk-1,iops_rd=5000,iops_wr=600,mbps_rd=50,mbps_wr=50,size=50G) to mc11-6-local-ssd
2017-01-17 17:10:49,985: INFO: zfs(1) send/recv 17.2GiB data from 'mc11-5-ssd/vm-109-disk-1@proxmove-170117-171049' to 'mc11-6-ssd/vm-159-virtio0' (on mc11-6-local-ssd)
mbuffer: warning: No controlling terminal and no autoloader command specified.
cannot open 'mc11-6-ssd/vm-159-virtio0': dataset does not exist
cannot receive new filesystem stream: dataset does not exist
mbuffer: error: outputThread: error writing to <stdout> at offset 0x80000: Broken pipe
summary: 512 KiByte in 0.0sec - average of 488 MiB/s
mbuffer: warning: error during output to <stdout>: Broken pipe
When changed to rpool (the correct name), things worked:
2017-01-17 17:17:40,735: INFO: Begin copy of 'virtio0' (mc11-5-local-ssd:vm-109-disk-1,iops_rd=5000,iops_wr=600,mbps_rd=50,mbps_wr=50,size=50G) to mc11-6-local-ssd
2017-01-17 17:17:40,735: INFO: zfs(1) send/recv 17.2GiB data from 'mc11-5-ssd/vm-109-disk-1@proxmove-170117-171740' to 'rpool/vm-159-virtio0' (on mc11-6-local-ssd)
mbuffer: warning: No controlling terminal and no autoloader command specified.
in @ 362 MiB/s, out @ 362 MiB/s, 17.4 GiB total, buffer 0% full
summary: 17.4 GiByte in 52.4sec - average of 339 MiB/s
Connection to 10.101.1.45 closed.
2017-01-17 17:18:33,876: INFO: Volume transferring/conversion 100% complete!
We should check this before commencing any moves.
See if we can improve the .proxmoverc config by auto-generating/testing storage settings:
http://pve.proxmox.com/pve-docs/api-viewer/index.html
Is this still an issue?