Posts made by bsodmike

@olivierlambert I have a fedora workstation, which I could use to serve an NFS share. Hmm, I could try this to isolate the Freenas share?

@olivierlambert That makes sense. FYI I only got the VDI error when attempting delta backup via XO. xe vm-export has not provided any errors.

Just right now I was able to export a 35Gig XVA file, reimport using xe vm-import and the VM boots just fine.

All over the SAME NFS share that doesn't work with XOA. Hmm...

@danp Here's an example that I just ran of a manual export to a NFS share that I setup on the host,

[23:22 xcpng-host0-mgmt xcpng-manual-export]# xe vm-export vm=docker1 filename=/mnt/xcpng_data/xcpng-manual-export/docker1_`date +"%d%m%Y"`.xva
Export succeeded
[23:28 xcpng-host0-mgmt xcpng-manual-export]# pwd
/mnt/xcpng_data/xcpng-manual-export
[23:29 xcpng-host0-mgmt xcpng-manual-export]# ls -lah
total 8.7G
drwxr-xr-x 2 1001 1005   3 Dec 20 23:23 .
drwxrwx--- 9 1004 1005  12 Dec 20 23:16 ..
-rw------- 1 root 1005 36G Dec 20 23:28 docker1_20122020.xva
[23:29 xcpng-host0-mgmt xcpng-manual-export]# mount -t nfs
freenas-primary.intranet:/mnt/big-primary/xcpng_data on /mnt/xcpng_data type nfs (rw,relatime,vers=3,rsize=131072,wsize=131072,namlen=255,acdirmin=0,acdirmax=0,soft,proto=tcp,timeo=600,retrans=2,sec=sys,mountaddr=10.0.0.149,mountvers=3,mountport=727,mountproto=tcp,local_lock=none,addr=10.0.0.149)

@danp I tried both actually, but right now I've switched to install via source using https://github.com/ronivay/XenOrchestraInstallerUpdater

@danp Also FYI, XOA failing backups (to the NFS share mentioned above) should not create smaller files and go "green" marking the backup as successful.

These are all failed attempts on host0.

@danp Hi Dan, I am easily able to export to FreeNAS via xe vm-export. The only issue I'm facing right now is getting XOA to run backups to NFS (or even CIFS) shares.

Right now, I do not have any other hosts to backup to, but I fail to see the FreeNAS host I'm using as the failure point given above. Thoughts?

@olivierlambert The delta backup finally shows an error:

Duration: 18 minutes
Error: VDI_IO_ERROR(Device I/O errors)

@olivierlambert

• Delta backup, similar behaviour, only 8% and then it appears to be "stuck" for the time being.
• host1 (the older one) is running xcpng 7.6.0. The newer host0 is running 8.2.0

Both are however running xo-server 5.73.0 and xo-web 5.76.0

@shorian Thanks for your response. Please see my new thread https://xcp-ng.org/forum/topic/3971/unable-to-export-reliable-backups/1

I tried adding more vCPUs and even upped the RAM to XOA to 8gig. It starts to generate a backup but then dies out 1-5GB into a 50GB output XVA.

Ideas?

Hi all,

I've extensively tested the above and here's the approach taken. I'm running XCPng on two separate hosts,

• xcpng-host1
• xcpng-host0 (newer server).

I have been able to verify via iperf3, ~9+Gbit/s works just fine, when connecting to my primary Freenas server at 10.0.2.2. I have modified /etc/fstab/ to mount a CIFS connection

# Connect to freenas-primary over 10G interface
//10.0.2.2/xoa-data-backup       /mnt/xoa-data-backup              cifs    uid=0,credentials=/root/xoa-backup.cifs,iocharset=utf8,vers=3.0,noperm 0 0

I have also tried adding an NFS share via XOAs "New file system remote", with the mount point \10.0.2.2:/mnt/big-primary/xcpng_data and options vers=3,soft,proto=tcp

Starting a backup, always looks like this:

Notice there's a blip of a transfer.

Ultimately the backup is marked as successful (takes ages) and the file size is between 1-5GB.

If I try to re-import via XOA, it just fails (never succeeds with error "IMPORT_ERROR_PREMATURE_EOF()". A manual xe vm-import reports The VM could not be imported; the end of the file was reached prematurely.

However, via the xcpng host I am able to issue xe vm-export and generate ~40GB+ XVA backups.

Thanks.

@olivierlambert Hah, indeed it was! CIFS is totally unstable, but I was able to get NFS working with these configs:

mount -o vers=3,soft,mountproto=tcp [IP]:/share /mnt/nfs_share/backup.xva

@olivierlambert bit more info

[20:42 xcpng-host0 ~]# mount -t cifs
//10.0.x.x/linux_data on /mnt/linux_data type cifs (rw,relatime,vers=3.0,cache=strict,username=myuser,uid=0,forceuid,gid=0,noforcegid,addr=10.0.x.x,file_mode=0755,dir_mode=0755,soft,nounix,serverino,mapposix,noperm,rsize=1048576,wsize=1048576,echo_interval=60,actimeo=1)
[20:42 xcpng-host0 ~]# cd /mnt/linux_data
# it just hangs here.  I have to kill the SSH session and SSH back in

@olivierlambert here's another clue. If I mount my local CIFS share onto the XCPng host, after a while I cannot do

cd /mnt
ls -la
# it just hangs here.  I have to kill the SSH session and SSH back in

This only happens after mounting the CIFS share the very first time after boot.

Thoughts?

@olivierlambert Thanks your comments got me thinking, it could be the NFS mount connection is dropping out; trying a test now to mount it via fstab in XCPng.

@olivierlambert Are there any compatibility issues, i.e. say exporting a XVA with an older version of XCPng and trying to import with a newer version?

@olivierlambert it's definitely not the storage, will try a different XVA to isolate.

@olivierlambert Nothing out of the ordinary. Still 1%.

@olivierlambert It's showing 1% but not doing much beyond that.

@olivierlambert Hang on there's a sliver of progress now!