XigmaNAS

Posted: **02 Sep 2014 11:31**

Hi,

I am trying to replicate from one N4F box to another one, and after a few days of transfer, I get this:

receiving incremental stream of pool1/dataset2@auto-20140806-060000 into pool1/dataset2@auto-20140806-060000
received 11.1GB stream in 28604 seconds (406KB/sec)
receiving incremental stream of pool1/dataset2@auto-20140807-060000 into pool1/dataset2@auto-20140807-060000
Write failed: Broken pipe
warning: cannot send 'pool1/dataset2@auto-20140807-060000': Broken pipe
warning: cannot send 'pool1/dataset2@replica-20140807-121100': Broken pipe

The command I am using is:

nasserver1: ~ # zfs send -R pool1/dataset2@replica-20140807-121100 | ssh 192.168.3.243 zfs recv -vd pool1

Is this recoverable, can I just send the failed differences?
Or must I start over from scratch and hope it doesn't fail this time?

Thanks

Posted: **02 Sep 2014 12:35**

You need to tune ZFS to avoid RAM / memory problems, and test until you do not have broken pipe errors.

viewtopic.php?f=71&t=1278&p=34461

Posted: **20 Sep 2014 11:18**

Hi raulfg3,

Thanks, I will look into memory problems.

But I feel this is because these two NAS boxes are away from each other (different offices) connected by an OpenVPN tunnel.

Could latency / network problems cause the broken pipe?

My real question is, in the event of network lag or disruption, is send/receive resumable?

Thanks.

XigmaNAS

Write failed: Broken pipe

Write failed: Broken pipe

Re: Write failed: Broken pipe

Re: Write failed: Broken pipe