You are not logged in.
XSIBACKUP-PRO is not a real time, or close to real time, syncronization tool. If you want such feature, you should think about some distributed file system. If you use a small hammer to tear a concrete wall down, you'll soon complain about it, whereas if you use a demolition hammer to nail, you'll soon be dissatisfied too.
You are probably thinking: "as this tool works so well to backup VMs fast, I might use it for something different" => Wrong approach, XSIBACKUP-PRO is a backup tool, use it for your backups and don't try to force it beyond it's limits, as it's boundaries are those of it's own nature.
You can also put it the other way round. If you try to use a distributed file system to keep a real time backup, you will soon discover why that's not such a great idea.
In other order of things, If a OneDiff cycle is taking more than 4 hours, then something is not working as it should and you should stop and find what's going wrong.
If you are using --on-success and --on-error event handlers, then backups should never overlap.
Well, isn't something as "[b]nothing has changed[/b]" and "[b]Does placing the hosts into Vcenter make any difference?[/b]" a bit contradictory. That's not a little detail with no importance.
Of course it can make a difference, but that will depend on how you configure the thing.
If you have DRS enabled and your VMs are moved around without you knowing it, XSIBackup might be trying to backup a VM that is not registered to the host. So, start by reading this:
[url]https://pubs.vmware.com/vsphere-50/index.jsp?topic=%2Fcom.vmware.wssdk.pg.doc_50%2FPG_Ch13_Resources.15.6.html[/url]
And then use XSIBackup's options as required:
[b][url=https://33hops.com/xsibackup-help-man-page.html#disablevmotion]--disable-vmotion[/url][/b] might be particularly useful.
Windows OSs are radically Windows, they play in a different league.
In any case that is a vmkfstools error which normally indicates there's something wrong in the .vmdk disks themselves. On the other hand, there aren't any known compatibility issues, so...
Start by rebooting your host and if you still get that nasty error, try to think if those VMs have something in particular that could be causing the issue, namely: did you convert them from physical or some other format?, if so, what software did you use?, are they aligned to 1M block size?
Check the descriptor files (plain tiny .vmdk files), to verify to what -flat.vmdk files they point to.
Same as before. Look for previous errors. The name of the VM should be between the first brackets, but it's printing empty.
On top of that, the syntax of the error is:
[VM_NAME] error DIFQMSH3: data size mistmatch:But yours is different. What version are you using?
There's not header, so we don't know.
This is specialized support, please contact us to fix your issue, I don't think It will help others.
Well, you have some permissions errors as you say. When you install xsibackup with the install script, it sets permissions automatically. If you install manually, you just have to run this one-liner:
chmod -R xsibackup binThis is stated in the README.txt file.
At Server 3 you are missing execute permissions for the xsibackup-rsync binary, start by fixing that and let's see what happens, cause it's not reason enough for the .ERR-XXX... file not to have the right permissions either.
This could be a bug. We'll check it out and eventually open a post at the [b][url=https://33hops.com/forum/viewforum.php?id=13]XSIBackup bug tracker[/url][/b]
UPDATE 2018-03-27 18:00 UTC:
We have performed some tests from ESXi 5.5.0 to 6.5.0 so far. No trace of bugs, when the checksums don't match the error is correctly recorded and informed. We'll continue to make tests.
Both of you suffer from the same error. The .ERR-XXXX... file is not present, this can only mean it was deleted by other process, or that XSIBackup was not able to create it. Which user are you running the XSIBackup process under?, do you have some other prvious error in the log?
Follow the white rabbit called [b]no peer certificate available[/b], use a different e-mail server, use your own local SMTP without authentication, use a different e-mail server without SSL, reinstall ESXi...
You have multiple possible ways to fix or work your problem around.
Make sure the --exec=yes directive is at the end of the backup job string
You have a definitive message
[b]no peer certificate available[/b]
Have you assigned a hostname?
does it make sense in terms of being resolvable?
OpenSSL is not presenting a certificate, maybe it can't for some reason.
[url]https://stackoverflow.com/questions/24457408/openssl-command-to-check-if-a-server-is-presenting-a-certificate[/url]
[quote=sistemi]Damn, i thought i was so near to heaven, no way to make it working, what's wrong here?
All working from command line, test date formati is weekday-hh-mm, doing ps -c|grep xsibackp i see the /bin/sh ... etc..., but joust for a moment, nothing on the log or .ERR
if [ "$(date '+%u-%H-%M')" == "3-14-03" ]
then
# usb mount
esxcli storage nfs add -H 192.168.1.1 -v usb -s /USBDisk2/
/vmfs/volumes/datastore1/xsi-dir/xsibackup backupid=E1BAKtoUSB --exec=yes --backup-prog=xsitools:z --backup-how=hot --override=xsibakfilter --backup-type=custom --test-mode=false --use-smtp=1 --mail-to=xxxxxxxx --subject="test backup" --backup-point=/vmfs/volumes/usb/aaaaaaa --debug-info=yes --backup-vms=zzz
#usb umount
esxcli storage nfs remove -v usb
fi[/quote]
Get used to employ leading zeros in dates.
In other order of things, for goodness sake, don't use USB devices to backup your VMs. USB is a nice method to plug your key and copy your office documents, but it's not a serious mean to backup critical data. If you don't know what I mean, I bet you'll end up knowing why we are vehement about it ;-)
We don't know if you are referring to the original VM being backed up or the resulting VM.
Provided that you are talking about the resulting VM produced by XSIBACKUP-PRO, the normal behaviour is that the differential data transmitted inside the snapshot, is integrated or consolidated with the pre-existing data on the target VM.
[url]https://33hops.com/xsibackup-pro-onediff.html[/url]
Nevertheless, a set of checks is performed in the original and target VMs (Trivial Check) to verify they are exact copies. An optional full hash check is available through the [b]--certify-backup[/b] argument. Should something as gross as a differential snapshot not having been integrated at the destination VM, some very obvious errors would be showing up throughout the logs and e-mail reports.
In case you are talking about the VM being backed up, a snapshot called [b]xsibackupdiff[/b] is always present when the VM is engaged in a OneDiff mirror. It's the snapshot that holds the differential data between backup cycles.
You have totally rewritten our example, I'm sorry sir but we don't correct code here.
On the other side, you are missing some arguments in your XSIBackup job: [b]--backup-prog[/b] being the most important one.
On addition XSIBackup will expect the job to include a [b]--time[/b] argument when called from the cron, thus you should add the directive --exec=yes to your job
I suggest that you try your command in the command line before using it.
Since version 10.3.2 process management has changed to be less agressive with ongoing jobs. In any case, you fell into a backup job design flaw, you should be using backup job chaining to prevent this kind of situations. Use --on-success and --on-error arguments to launch subsequent backup only when the previous has finished.
Note: running two backups at the same time will not last as the sum of running them subsequently, it will take much, much longer, and only if you are lucky to have decent hardware, cause you might very well clogg your server up to a point in which it just says "I had enough". We won't delve into the details of why this is as is, although many of you have that clear.
The bug this thread started with was resolved some time ago and had nothing to do with Rsync.
Please open a new thread and describe your problem in detail. We cannot help you if you don't help us to help you first.
We currently perform backups on suspended VMs without issue. Maybe you have a different type of problem, your log above shows, at least, one disconnected datastore.
Disk exclusions don't work on 10.3.0 free, this issue has been solved in XSIBACKUP-PRO 10.3.2
There's a much more slick solution, which we overlooked. In fact we've been about to add it to the skeleton xsibackup-cron file a number of times and we forgot, we'll add it today, so that it is available with next version.
The xsibackup-cron file is a bash file, which is called every minute from the ESXi crontab. You can't add there whatever you want, cause it would execute each minute, but that has a simple solution.
if [ "$(date --date="2018-03-16 16:33" +%Y-%m-%d'T'%H':'%M )" == "$(date +%Y-%m-%d'T'%H':'%M )" ]
then
echo BINGO!!
fiBy using this simple technique, you can program whatever other task and the crontab will be persistent across reboots.
It's just a new version availability check. If you don't have proper Internet access, it may take some time to timeout. To work around this situation, just change the var [b]checkver=yes[/b] to [b]checkver=no[/b] at the config file [b]conf/xsiopts[/b]
[url]https://33hops.com/xsibackup-pro-onediff.html[/url]
Enable "Less Secure Apps" for your GMail account or use a regular e-mail account.
[url]https://33hops.com/xsibackup-smtp-client.html[/url]
We know that, it's a current XSITools limitation that will be improved in future versions. By now, if you want to limit the size of XSITools repositories, you should not backup VMs containing big snapshots.
Production machines don't usually work on top of snapshots, except if they are engaged in a OneDiff chain per instance. If for whatever reason your VM needs to keep big files of this kind, just make a regular backup (vmkfstools or onediff) and then archive the resulting VM with XSITools.
[url]https://33hops.com/xsibackup-pro-onediff-backup-retention-policy.html[/url]
If you need your backup to contain the snapshots as they are in the original VM, then use --backup-prog=xsidiff
We can offer you some ideas, but you'll have to implement your working solution.
I guess your script carries all the logic: mount NFS, xsibackup and unmount NFS. Provided that you have tested your script and it's working, the only thing left to do is scheduling, right?
You then have multiple choices:
[b]1 -[/b] ESXi cron, as you said, this is cumbersome to implement, as the ESXi crontab is not persistent, god knows why.
[b]2 -[/b] Use an external Linux cron. When I say external I mean external to ESXi, but this could very well be a VM in the same ESXi host. Just exchange keys manually and program an SSH call to your script. Do not forget to use -t(t) SSH's option and/ or nohup as needed. You have a guide here: [url]https://33hops.com/xsibackup-cron-how-to.html[/url]
[b]3 -[/b] Use XSIBackup cron. It has not been designed to call bash scripts, it calls itself, but as XSIBackup will keep on rolling, no matter what errors it encounters, you just need to call XSIBackup with a malformed job syntax, like no --backup-point, and then use the [b][https://33hops.com/xsibackup-help-man-page.html#onerror]--on-error[/url][/b] and [b][https://33hops.com/xsibackup-help-man-page.html#onsuccess]--on-success[/url][/b] options to run your script.
UPDATE:
[b]4 -[/b] I can think of a fourth option, which is use a Windows command line SSH client, like [b]plink.exe[/b] bundled with Putty. Create a batch, VB Script or Power Shell script (a batch is more than enough for the case) that runs your remote script and finally add your local Windows script to your task manager. Other Windows options are using CygWin ([url]https://www.cygwin.com/[/url]) or the embeded Windows 10 Linux Shell ([url]https://www.howtogeek.com/249966/how-to-install-and-use-the-linux-bash-shell-on-windows-10/[/url])
No there isn't any way and I would not do that if you appreciate your data, just let XSITools do it's job.
[url]https://pubs.vmware.com/vsphere-51/index.jsp?topic=%2Fcom.vmware.vsphere.storage.doc%2FGUID-8AE88758-20C1-4873-99C7-181EF9ACFA70.html[/url]
If you run tests in the console and you Ctrl+C, you may leave some process waiting.
Locate any eventually running XSIBackup process
youresxi# ps -c | grep xsibackupAnd kill all XSIBackup loaded processes if you need to.
kill -9 $(ps -c | grep xsibackup | awk '{print $1}')