You are not logged in.
[b]Failed to clone disk: Input/output error[/b]
The above is a low level transport error which undoubtedly means your NFS share is not working as expected. It's an error raised by [b]vmkfstools[/b] when it can't read or write data from the share.
The fact that you can browse or create some file on the share does not mean you can copy a gigabyte file. Try to copy the same file manually and you will, most probably, receive the same error. Backup to some other local datastore and you will most probably not get any errors.
The cause for this could be from some bad cable to some inconsistency in the NFS protocol configuration. Do make sure that you are not connecting as NFS 4.1, as this is the most common cause for this kind of errors, apart from physically damaged cables or disks.
Then inspect your descriptor files as stated above and find out where those [b]-flat.vmdk[/b] files are, cause if there are no [b]-flat.vmdk[/b] files, there is no data, there are no VM.
View the contents of your descriptor file, the small one that ends in .vmdk, not -flat.vmdk and see to which -flat.vmdk file it is pointing and whether it is the correct one.
We don't support manufacturer builds. They are plainly not ESXi, as they are heavily customized and they frequently contain bugs.
That does not mean that [b](c)XSIBackup[/b] will not work with manufacturer builds but that we can guarantee nothing, as manufacturers frequently do not respect basic principle designs of ESXi.
Your problem seems related to your .vmdk files having been edited by some non UTF-8 text editor or containing bad pointers to -flat.vmdk files.
You will have to skip 1st cycle then. Set the OneDiff backup and allow VM auto registering, then start XSITools from the second cycle.
We know that, we may change that behaviour in future versions, but that shouldn't make a difference in any case.
[b]--options=unreg-xsibak[/b] will unregister the VM in every backup cycle. This is convenient if you are in a [b]vCenter[/b] context, as it might check for virtual MAC addresses duplicates even on switched off VMs.
Register the VM manually if you need to use it and try not to register VMs automatically to prevent MAC duplicates.
Never mind, just as long as it's useful to other people...
Yes, you can use vi to edit the job files or any other XSIBackup related file, nevertheless, the undeniable fact is that what you pasted has been changed the page code ([b]^M[/b] termination). Whether it's been in the paste process or before, that's something we have no epistemologic access to.
Check whether those line termiinations are present at your [b]xsi-dir/conf/root-crontab[/b] when visualizing it on vi. If so, I would not bother much trying to find out how or denying the fact that they are there, just reinstall again.
Pro users may write to support to request a working version, which basically consists in XSIBackup running on its own shell. Free users should not worry much, just wait for VMWare to fix the issue.
ESXi is still the best hypervisor by far, but this guys at VMWare are going downhill lately, you can tell a "disturbance in the force" at their tech HQ ;-)
You are posting to [b]XSIBackup-Pro[/b] forum, so if you are using Free version, this is not the place to post your questions.
You have edited the job files in an external editor and you have broken its page code, you can tell by the [b]^M[/b] at the end of every line.
VMWare released a patch for this bug.
[url=https://33hops.com/forum/viewtopic.php?pid=2104#p2104](c)VMWare patch for the Bash interpreter[/url]
Nevertheless, for goodness sake!, don't use in production some OS that just came out...
We sometimes use metaphores, gags or funny stories like being the "Caesars food taster", but this is something serious.
Nobody with some minimum experience and know how will put into production some OS that has been released "the day before".
Use some stable and well known version, most stable ESXi version from our experience is 5.5, although 6.0 and 6.5 have been out long enought to have some guaratees about its stability too.
This is a general comment, if you were just testing 6.7.0 U2, then this is obviously not for you.
VMWare has finally released a fix for the broken bash in ESXi 6.7.0 U2
[url=https://docs.vmware.com/en/VMware-vSphere/6.7/rn/esxi670-201904001.html#esxi670-201904401-bg-resolved](c)VMWare patch for ESXi 6.7 U2 Bash interpreter[/url]
We recommend that you skip this version until a more stable build is reached or that you patch your ESXi box with the most recent fix to use XSIBackup.
Well, there isn't really something that could be considered a "complex install". ESXi installs in some minutes. If you have some config backup (XSIBackup does them by default), you just need to reinstall ESXi preserving your datastores, then just unzip the contents of the server config backup and overwrite the /etc directory, that's all.
You can also use the built in commands. Copy the config backup to the temp dir and use the vim-cmd utility to restore it:
vim-cmd hostsvc/firmware/restore_config /tmp/backup-file.tgzIt's a pretty straight procedure that works like a charm. Just as long as you install the same ESXi version that the backup corresponds to, you won't have any problem.
Copy [b]XSIBACKUP-FREE.zip[/b] to you /tmp folder and install from there.
Remove the current xsi-dir directory before proceeding.
This entry in your [b]root-crontab[/b] is wrong:
58 12 * * * "/vmfs/volumes/Areca_Raid10-04_DEV01/xsi-dir/jobs/010" > /vmfs/volumes/backup/log.log 2>log2.logIt should be:
58 12 * * * "/vmfs/volumes/Areca_Raid10-04_DEV01/xsi-dir/jobs/010"Use the GUI to generate the contents of your [b]root-crontab[/b] file by going to "Manage backup jobs" => "Schedule"
Why don't you just use a more stable ESXi version: 5.5, 6.0, 6.5, 6.7 U1
You can't just download a binary and put it into your ESXi server. Please download latest version to fix your issue.
Well, we have not tested Sistemi procedure, so anyone trying this out does at his own risk.
We have developed a new version that incorporates a fix as XSIBACKUP-PRO 11.3.0
The recommended procedure is:
1 - Skip ESXi 6.7.0 U2 if you can.
2 - Contact support to get your 11.3.0 version copy.
We have not published it yet as it's in testing phase, it's working fine so far anyway.
For those of you using more intrincate topologies, some bug may still arise.
There's what seems to be a bug in the bash interpreter shipped with ESXi 6.7 U2. We have developed a solution as XSIBackup-Pro 11.3.0, but it's still in validation phase. In this post we treat the matter and offer a workaround.
We recommend that you avoid ESXi 6.7.0 U2 until VMWare releases a patch. If you can't skip this build and you are a Pro user, you can try XSIBACKUP-PRO 11.3.0, it's working quite well for us so far. Write to support to ask for it.
You need the Pro version to launch remote jobs via the [b][https://33hops.com/xsibackup-help-man-page.html#manhost]--host[/url][/b] option. Yes, you can use any outer cron system that can reach your ESXi servers (your fridge if it can run crond), you can put the public SSL key of that cron scheduler in each of the ESXi servers' [b][url=https://www.ssh.com/ssh/authorized_keys/]authorithed_keys[/url][/b] file to allow it execute jobs without login, or you can use any SSH client that allows implicit password authentication, like [b][url=https://linux.die.net/man/1/sshpass]sshpass[/url][/b]
Rename the job:
mv 2 002And change the name of it in your cron files:
- [b].../xsi-dir/root-crontab[/b]
- [b]/var/spool/cron/crontabs/root[/b]
You may need to change permissions on your ESXi crontab to be able to edit it, if so:
chmod 0700 /var/spool/cron/crontabs/root
vi /var/spool/cron/crontabs/root
chmod 0600 /var/spool/cron/crontabs/rootYes, of course. Job Ids are strings, not numbers, thus 2 != 002
We can provide an upgrade (v. 11.2.5) that contains a fix for this issue. The fix consists in XSIBackup being shipped with its own bash interpreter, still XSIBackup will depend on the other Busybox utilities present in ESXi PATH variable.
Contact support to get the update, we won't publish it for download ultil it has gone through the regular verification and testing process.
The near future evolution of XSIBackup will be to use its own shell, avoiding this way any incompatibility caused by the ESXi shell utilities. It will then depend on the ESXi's GlibC library mainly.
Keys generated by [b]ssh-keygen[/b] are not recognized by the OpenSSL binary.
openssl rsa -in my_ssh-keygen_generated_key -check
unable to load Private Key
331484505768:error:0906D06C:PEM routines:PEM_read_bio:no start line:pem_lib.c:697:Expecting: ANY PRIVATE KEYThe key to this issue is that there exist newer methods for key generation by using the openssl base binary ([b]openssl genrsa[/b]). Checking of keys does work when using this method, thus [b]ssh-keygen[/b] binary should have been removed, as it's not compatible any more.
We haven't delved much into this, but it all seems to be some sort of incompatibility about the key file format, maybe the encoding. The error is the same that would be thrown by the -check method if we tried to parse any file other than a key.
Yes, job files aren't supposed to hold comments. You have a description argument for that purpose.
Using the latest software in production is playing Russian roulete with your feet. This case is a clear example of that statement.
Some changes were made in the Bash interpreter that would break Bash script execution.