@stormi I've placed the tarballs here https://student.oedu.se/~peder/xcp-ng/
I changed the static min to 512M, to match the Fedora case, but it still failed.
Olivier, I'm not using Xen Storage motion but I am using two old Lenovo L430 Thinkpads as "servers" so that could be part of the problem.
I'll install a new C7 guest and see if the problem persists.
@dan2462 Don't know what's causing the error but you can export the VHD by ssh:ing to your server or selecting "Local Command shell" on the console and running the following
xe vdi-list
xe vdi-export uuid=the-uuid-of-your-VM filename=filename.vhd format=vhd
Make sure the location you're saving to has enough space for the file.
If you have curl installed on your VirtualBox machine you can run
curl -u root:your-xcpng-root-password -X GET "server-name-or-IP/export_raw_vdi?vdi=your-vdi-uuid&format=vhd" -o filename.vhd
on it.
To run a command in the current directory you have to prepend it with ./
So try ./recover-account-cli.mjs
I don't know if it's the case here but one not-so-obvious thing in XO(A) is that you have to enable the slider to accept unverified certificates when you try to connect to the XCP-ng server.
The error here does suggest there's something else going on but make sure to check the above.
You should also install the XOA to make sure there's nothing wrong with your self-compiled XO.
Run this in a command shell on your XCP-ng server to install the appliance
bash -c "$(curl -s http://xoa.io/deploy)"
You can also run the vm-clone command on the XCP-ng server.
Something like
xe vm-clone new-name-label="clone name" vm="original name"
Stopping xo-server before compiling shouldn't be necessary unless someone tries to use XO while it's building.
That said, I would recommend building the source under /usr/src or ~/src or something and then moving the files ( mv * ) to something like /opt/xen-orchestra-DATE (or git revision)
Then use "stow" to create symlinks in /usr/local, like this
stow -d /opt/ -t /usr/local xen-orchestra-DATE
The stow command is in the CentOS repository or can be downloaded and compiled separately from any GNU mirror.
Make sure the /etc/xo-server/ config points to the relevant symlinks in /usr/local and restart xo-server
This way you can keep a number of different XO versions available in /opt and if there's a problem with any one of them you simply run
stow -D -d /opt -t /usr/local xen-orchestra-BADVERSION
to delete the symlinks to the bad version and
stow -d /opt -t /usr/local xen-orchestra-GOODVERSION
to create links to the old, good one and then restart xo-server
When it's time to update the XO build go to the, now seemingly empty, source folder and run
git reset HEAD --hard
to repopulate the working directory and then
git pull && yarn && yarn build
AFAIK XO can only create a new storage repository, it would be nice to being able to attach an existing one as well.
@gthvn1 Yeah, that works.
I can get the logs both from the same host and from the other one in the pool.
I managed to download from master now as well, with the same gzip complaint and the same "Forkhelpers.Subprocess_failed(1)" in the log
@andriy.sultanov I'm not seeing any errors after it stalls.
One funny thing is that the pool master usually stalls after approximately 135MB and the slave stalls after 230MB
/var/log on master is half the size of slave's so that may be why.
Another funny thing is that when I tried the slave once more it actually finished the download at 230MB but my subsequent "tar -xf" complained about "gzip: stdin: decompression OK, trailing garbage ignored"
And for some reason the files it extracted were 528MB whereas they're 498MB on the server.
Maybe that's a matter of my using XFS vs the server's ext3, though.
That "successful" download resulted in
xapi: [error||307417 :::80||backtrace] Xapi#getResource /host_logs_download R:51bd50ee890d failed with exception Forkhelpers.Subprocess_failed(1)
in xensource.log.
So it seems to fail to properly close the transfer session most of the time and the file compression seems to be somewhat buggy as well.
Maybe those things are related somehow.
@stormi Any hint on what to look for?
I see these entries when the download starts but no subsequent errors or anything else that stands out when the download stalls
xapi: [ info||1520711 HTTPS 10.x.y.z->:::80|task.create D:b06a0108fbb7|taskhelper] task Xapi#getResource /host_logs_download R:8ea91ee79d9a (uuid:a-b-c-d-e) created (trackid=18e32f90af09a295c704ca6099aaef26) by task D:b06a0108fbb7
xapi: [debug||1520711 :::80||xapi_logs_download] running logs-download handler
xapi: [ info||1520711 :::80||taskhelper] task Xapi#getResource /host_logs_download R:8ea91ee79d9a forwarded (trackid=18e32f90af09a295c704ca6099aaef26)
xapi: [debug||1520711 HTTPS 10.x.y.z->:::80|Xapi#getResource /host_logs_download R:8ea91ee79d9a|xapi_logs_download] send the http headers
Same problem here. In my case it downloads about 30MB before it stalls.
XCP-ng version : 8.2.1 with March updates
XO : I tried both 9c27b53df and the older 708b9b5bc
Browser : Firefox and Chrome in Linux
To run a command in the current directory you have to prepend it with ./
So try ./recover-account-cli.mjs
@AlexQuorum Does
openssl x509 -noout -text -in /etc/xensource/xapi-ssl.pem | grep bit
show 2048 or more and did you remember to do a
systemctl restart xapi
after you replaced the cert?
@tony Is your storage thick or thin provisioned, if you click on the Advanced tab?
If it's a thin storage it could be that the amount "used" is the disk size you've specified when creating the VMs and the bar represents the actual number of GBs used right now.
@saneece said in XO Lite: building an embedded UI in XCP-ng:
Can you explain why the calling on HOST_URL/xolite.html did not work and why I needed your special URL
Yeah, I was just explaining why HOST_URL/xolite.html didn't work for you.
You only ran wget https://lite.xen-orchestra.com/
and that's why your log screenshots says that it's saving to index.html.
If you had run the entire command it would have said "Saving to /opt/xensource/www/xolite.html " and you would have been able to access it the "normal" way.
@saneece The XO Lite installing instructions say that you should run this command on your xcp-ng server
wget https://lite.xen-orchestra.com/ -O /opt/xensource/www/xolite.html
You seem to have omitted the -O /opt/xensource/www/xolite.html part, since your screenshot talks about "Saving as index.html", and that's why you can't go to HOST_URL/xolite.html