[SOLVED] "Error '0' occured while receiving the doument" when uploading an iso as pve user

[BerryJP]

I have the same issue with a TrueNAS ISO.

I'm going to try the fix now.

UPDATE: Didn't work. Looking for other fixes now.
UPDATE 2: It only worked when I accessed the webUI @ the IP of the node I wanted to upload to in the cluster.

 

[loveablecabbage]

Yes, that's a possible reason.
I currently can't reproduce the issue, but you could try fixing it yourself by finding the following line in /usr/share/perl5/PVE/APIServer/AnyEvent.pm

$reqstate->{hdl}->on_read(sub {

and insert the line $reqstate->{hdl}->timeout(60); so it looks like that :

$reqstate->{hdl}->timeout(60);
$reqstate->{hdl}->on_read(sub {

This sets the server side timeout to 60 seconds, which should be plenty.

Execute systemctl restart pveproxy.service.
Now upload the file.

If you do, please report back whether that fixes the issue.

To undo the changes, run apt install --reinstall libpve-http-server-perl

This worked for me as well, thank you. PVE 7.4-3

 

[jacobtheeldest]

I am experiencing the same issue as others in this thread on PVE 8.0.3. Uniquely, I am using the bpg/proxmox terraform provider, but it is using the API under the hood, so I am fairly certain I'm in the same boat.

I am attempting to download an image from this URL: https://cloud-images.ubuntu.com/jammy/current/jammy-server-cloudimg-amd64.img

My client is giving me a 502 error from the reverse proxy:

2023-06-28T19:39:22.150Z [ERROR] vertex "proxmox_virtual_environment_file.ubuntu_cloud_image" error: error uploading file to datastore local: received an HTTP 502 response - Reason: Bad Gateway

And my reverse proxy (nginx) is giving me upstream errors:

2023/06/28 15:39:21 [error] 590#590: *13424 SSL_write() failed (32: Broken pipe) while sending request to upstream, client: 10.0.0.1, server: pve.*, request: "POST /api2/json/nodes/pve-ws/storage/local/upload HTTP/1.1", upstream: "https://10.0.1.11:8006/api2/json/nodes/pve-ws/storage/local/upload", host: "pve.redacted.com"

Following the path up to the proxmox node gives me the following associated line in syslog:

Jun 28 15:39:21 pve-ws pveproxy[23833]: problem with client ::ffff:10.0.1.200; No space left on device

Like others, I have plenty of free space and inodes:

$ df -h
Filesystem      Size  Used Avail Use% Mounted on
tmpfs           745M  2.4M  743M   1% /run
/dev/sda1        31G   21G   11G  68% /
tmpfs           2.7G     0  2.7G   0% /dev/shm
tmpfs           5.0M     0  5.0M   0% /run/lock
/dev/sda15      105M  6.1M   99M   6% /boot/efi

$ df --inodes -h
Filesystem     Inodes IUsed IFree IUse% Mounted on
tmpfs            931K  1.2K  930K    1% /run
/dev/sda1        4.0M  530K  3.5M   14% /
tmpfs            687K     1  687K    1% /dev/shm
tmpfs            687K     3  687K    1% /run/lock
/dev/sda15          0     0     0     - /boot/efi

I did try adding the timeout in /usr/share/perl5/PVE/APIServer/AnyEvent.pm, but did not see any change in behavior.

Is this issue still being tracked?




Potentially related threads:
https://forum.proxmox.com/threads/image-upload-fails-after-upgrading-from-7-1-to-7-3.119051/
https://forum.proxmox.com/threads/storage-rest-api-error-no-space-left-on-device.117296/
https://forum.proxmox.com/threads/pveproxy-no-space-left-on-device-problem-with-storage-api.113451/

 

[Chlorgas]

The same happens on my machine while uploading to a M2.NVMe-Mirror on PVE 8.2.2.
So I want to repeat the question: Ist this issue addressed in any way?

 

[janthiel]

Hi, we have the same issue as root. No ISO upload possible. Our PVE version is 8.2.4.

 

[rnldnkp]

Same issue here (even with root user).

"Jul 25 21:49:09 pve pveproxy[218604]: problem with client ::ffff:10.x.x.x; Connection timed out"
(timeout/error in browser takes place in like 2-3 seconds)

Upload not working, tried different browsers, client OS, ISO OS. Disk only has 2% in use (500 GB disk)

Reading this topic I thought it had to do with my upgrade to 8.2.4.
But then I realized I first setup PVE using a 2.5 GbE Nic to test it. After a week I changed to a 10/25 GbE NIC in a different VLAN/different IP. About the same period I upgraded. I changed the IP, Gateway and host file record. Even generated new certificates. I am using a (VLAN Aware) Linux bridge + Linux VLAN interface with VLAN ID to set the IP/GW of PVE. Sadly ISO upload is still not working.

However! Once using the old IP (which is stil connected) I can upload like before.

Might help other although I have no clue how to fix this on the new interface/IP.

--

Nevermind.
Somehow there was a 'bogus' bridge still active which seemed to get things mixed up... (no other explanation)
This somehow forced all traffic trough the 2.5GbE interface, even though there was no gateway set.
So when I shut the 2.5 GbE port on the switch both IP's/NIC would not respond to ping even though the 10 GbE was ip with IP/GW.

Had to use brctl to remove the interface from the bogus bridge (not shown in PVE) and reboot. (removing bridge command kept failing)
Afterwards the interfaces seem to work as it should and the 10 GbE interface/IP is able to upload an ISO

Pff, long story.

 

[Ľudovít Lučenič]

Hi, we have the same issue as root. No ISO upload possible. Our PVE version is 8.2.4.

Hello, I have solved the problem with 'no space left on device' by setting correct permissions:

> chgrp www-data /var/tmp
> chmod g+w /tmp/var

The problem persisted even after upgrade to 8.2.3 and now it is gone. Not even necessary to add the timeout code line in Perl source.