Very good. I will try it later. I have solved it a different way now.
I have put the machine on Folding.

But Rosetta will be worth a try again on another machine when I can start up a Ryzen 3900X in a couple of months.
You effort is quite remarkable.

---

VERR_ALREADY_EXISTS tells you there's a disk registered with this name/UUID

Try
vboxmanage list hdds

But to answer the question, here is what it shows:

UUID: ef35dff9-d482-48f8-9519-fef6c1b23a3b
Parent UUID: base
State: inaccessible
Type: normal (base)
Location: /var/lib/boinc-client/projects/boinc.bakerlab.org_rosetta/AIMNet_vm_v2.vdi.bak
Storage format: VDI
Capacity: 0 MBytes
Encryption: disabled

UUID: f93b7440-2207-4969-bc1f-bbd87433063a
Parent UUID: base
State: inaccessible
Type: normal (base)
Location: /var/lib/boinc-client/projects/boinc.bakerlab.org_rosetta/AIMNet_vm_v2.vdi
Storage format: VDI
Capacity: 0 MBytes
Encryption: disabled

I will deal with it later.

EDIT: I closed them both, in preparation for a new try later, though this machine is really better on Folding anyway, and I will probably use the 3900X.

Thanks loads.

---

Will <open_name>vm_image.vdi</open_name> break something or does it just take hard drive space?

A "file" must be there (in the slots directory).
Beside that, in this special case the "file" has not other function any more.


With "<open_name>vm_image.vdi</open_name>" the file is a "softlink" called vm_image.vdi and pointing back to AIMNet_vm_v2.vdi in the project directory.
BOINC doesn't care but it may confuse users looking through the slots directories.

Without "<open_name>vm_image.vdi</open_name>" the softlink file will be called AIMNet_vm_v2.vdi.


In both cases the softlink file's size is just a few bytes.

Still download error: Rosetta@home | [error] Application file AIMNet_vm_v2.vdi missing signature

I did:
- Extracted AIMNet_vm_v2.vdi from AIMNet_vm_v2.vdi.gz
- Renamed "AIMNet_vm_v2.vdi" to "AIMNet_vm_v2.vdi.bak"
- Ran "vboxmanage clonemedium AIMNet_vm_v2.vdi.bak AIMNet_vm_v2.vdi" (Important: this creates a new UUID for AIMNet_vm_v2.vdi)
- Disabled the client's network access
- Resetted the Rosetta project (this removes the app_version you got from the project; must be done to avoid getting mixed versions)
- Shutdowned the client
- Created app_info.xml
- Modified vbox_job_v4.xml
- Copied the files: vboxwrapper_26205_windows_x86_64.exe
AIMNet_vm_v2.vdi (the already cloned one)
app_info.xml
vbox_job_v4.xml (the already modified one)
boinc_app_v4
- Restarted Boinc Client
- Enabled client's netword access

What am I doing wrong?

---

Post the app_info.xml.
Is it located in the Rosetta projects dir?
Does your BOINC client log mention an app_info (or anonymous platform) is loaded?

Are there any <file> or <file_signature> tags in the Rosetta section of client_state.xml before you restart the BOINC client?

Post the app_info.xml.

For sure!

<app_info>
    <app>
        <name>rosetta_python_projects</name>
        <user_friendly_name>rosetta python projects</user_friendly_name>
        <non_cpu_intensive>0</non_cpu_intensive>
    </app>
    <file_info>
        <name>vboxwrapper_26205_windows_x86_64.exe</name>
        <executable/>
    </file_info>
    <file_info>
        <name>boinc_app_v4</name>
        <executable/>
    </file_info>
    <file_info>
        <name>vbox_job_v4.xml</name>
    </file_info>
    <file_info>
        <name>AIMNet_vm_v2.vdi</name>
    </file_info>
    <app_version>
        <app_name>rosetta_python_projects</app_name>
        <version_num>103</version_num>
        <avg_ncpus>1.000000</avg_ncpus>
        <plan_class>vbox64</plan_class>
        <api_version>7.7.0</api_version>
        <file_ref>
            <file_name>vboxwrapper_26205_windows_x86_64.exe</file_name>
            <main_program/>
        </file_ref>
        <file_ref>
            <file_name>boinc_app_v4</file_name>
            <open_name>shared/boinc_app</open_name>
            <copy_file/>
        </file_ref>
        <file_ref>
            <file_name>vbox_job_v4.xml</file_name>
            <open_name>vbox_job.xml</open_name>
        </file_ref>
        <file_ref>
            <file_name>AIMNet_vm_v2.vdi</file_name>
        </file_ref>
        <dont_throttle/>
    </app_version>
</app_info>

Is it located in the Rosetta projects dir?

Yes

Does your BOINC client log mention an app_info (or anonymous platform) is loaded?

Yes

Are there any <file> or <file_signature> tags in the Rosetta section of client_state.xml before you restart the BOINC client?

This is the part of <file_signature> tags now, after the restart.
I haven't looked before.

<scheduler_url>https://bwsrv1.bakerlab.org/rosetta_cgi/cgi</scheduler_url>
    <code_sign_key>
1024
da94985671f399f2ccbb8711004a4d7b782f239babc54e4db341dd1c7b94fbf3
745d30084d332667546f400f5063e683c812a69a5d0945f53f0421961337e3f0
cfad19652eb4f50523473f92ee3b1f43d358a5ba911479e553f43c91b8a4939a
6aa5258107ef609a240bcffcfc9a19c8a8b0df99fdb9508694d499478fb0a931
0000000000000000000000000000000000000000000000000000000000000000
0000000000000000000000000000000000000000000000000000000000000000
0000000000000000000000000000000000000000000000000000000000000000
0000000000000000000000000000000000000000000000000000000000010001
.
</code_sign_key>

---

Looks fine.

You may add this option in cc_config.xml and try it again:
<dont_check_file_sizes>1</dont_check_file_sizes>

You may add this option in cc_config.xml and try it again:
<dont_check_file_sizes>1</dont_check_file_sizes>

I have no cc_config.xml.
Have i to create it?

---

I just checked the RPC replies.
Looks like the project sends the vdi info with each WU.
This means that you need a modified (development) client to get this working.

<edit>
... or a project that behaves more friendly, like LHC.
</edit>

<edit>
... or a project that behaves more friendly, like LHC.
</edit>

That is my conclusion too. I am trying to get the new CMS on the .dev site.
(I know, a bit off topic, but nobody knows what the real topic is here anymore).

---

<edit>
... or a project that behaves more friendly, like LHC.
</edit>

That is my conclusion too. I am trying to get the new CMS on the .dev site.
(I know, a bit off topic, but nobody knows what the real topic is here anymore).

Here in the UK , topic used to be a chocolate bar with hazelnuts in it :-)

Today Boinc developers released a new official version of boinc wrapper.
But i don't know if Rosetta team will be interested

---

Right.
:-)

It's now up to the Rosetta developers to make use of the multiattach feature.
Just in case they don't know where to download the required vboxwrapper.

Apple:
https://boinc.berkeley.edu/dl/vboxwrapper_26206_x86_64-apple-darwin.zip

Linux:
https://boinc.berkeley.edu/dl/vboxwrapper_26206_x86_64-pc-linux-gnu

Windows:
https://boinc.berkeley.edu/dl/vboxwrapper_26206_windows_x86_64.exe.zip

Hoping that 2023 brings these updates

---

Very interesting talk of computezmle about vbxwrapper at Boinc Workshop

---

Some virtualbox wus on Ralph.
I crunched only 1 so i don't know if they implemented the multi-attach funtionality....

---

I crunched only 1 so i don't know if they implemented the multi-attach funtionality....

Today, 5 wus, over 76Gb of disk space used, so no multi-attach disk.
Please, please, use this funcionality!!

---

I also suggest to switch the VM's virtual disk adapter to sata/IntelAHCI (via vbox_job.xml) and to get rid of forcing a VM snapshot every 10 min.
See:
https://boinc.berkeley.edu/trac/wiki/VboxApps

From what i know, the wiki pages are no more updated/manteined
See this for update situation of vbox wrapper

---