Cloudron Forum

I've been trying to use streaming within N8N on a ai agent chat node. I've enabled the streaming options on both the chat and ai agent, but it only ever seems to use the method of sending the full completion when the process finishes.

The logs only seems to point out the following message when I use it, but I don't think it's related.
Sep 30 07:41:49 Custom data value over 512 characters long. Truncating to 512 characters.
Sep 30 08:02:51 Custom data value over 512 characters long. Truncating to 512 characters.

Is there something more I need to do here to get streaming within the chat node since this is running in cloudron?

@nebulon

Hi, it's been a little while since this was last looked into. With the suggestion from robi do you think this is something we can look at doing soon? I am still willing to provide access for testing CUDA.

@nebulon I can provide access to mine if you think that will help to implement it? I'm sure a fair amount of people would love to use CUDA features with it.

@nebulon

It was never using local backups to begin with.


it does still have some data, but 200 gigs, it does not.

I was able to get into the docker folder and check to see what was taking up the space. It was all focused on one folder volumes/d2e397a84d2e73d5a8ce97860a01e4d6c4929e6a25fcff0af02c316b0a82d981/_data/

It seemed to contained large files that I transferred onto the server to then move somewhere else. I used cubby for this, so that makes me wonder if cubby retains what is uploaded to it in the docker folder for some reason?

For my uploads I would upload to cubby, then go into the file manager of cloudron to cut paste where I wanted the files.

I don't know if that is a bug or not but I at least found the cause of it by removing everything in there.

I am really baffled by this problem I am having. Every email from gmail to the inbox can't be received because of a timeout after it's trying to upgrade the tls

Oct 04 09:24:17 [INFO] [B2ED6280-367A-4095-95EE-B6D2484234B8] [spf] scope: helo, result: None, domain: mail-pl1-f171.google.com
Oct 04 09:24:17 [DEBUG] [B2ED6280-367A-4095-95EE-B6D2484234B8] [core] hook=ehlo plugin=spf function=helo_spf params=mail-pl1-f171.google.com retval=CONT msg=""
Oct 04 09:24:17 [DEBUG] [B2ED6280-367A-4095-95EE-B6D2484234B8] [core] running ehlo hook in limit plugin
Oct 04 09:24:17 [DEBUG] [B2ED6280-367A-4095-95EE-B6D2484234B8] [core] hook=ehlo plugin=limit function=max_errors params=mail-pl1-f171.google.com retval=CONT msg=""
Oct 04 09:24:17 [DEBUG] [B2ED6280-367A-4095-95EE-B6D2484234B8] [core] running capabilities hooks
Oct 04 09:24:17 [DEBUG] [B2ED6280-367A-4095-95EE-B6D2484234B8] [core] running capabilities hook in tls plugin
Oct 04 09:24:17 [DEBUG] [B2ED6280-367A-4095-95EE-B6D2484234B8] [core] hook=capabilities plugin=tls function=advertise_starttls params="" retval=CONT msg=""
Oct 04 09:24:17 [DEBUG] [B2ED6280-367A-4095-95EE-B6D2484234B8] [core] running capabilities hook in cloudron plugin
Oct 04 09:24:17 [DEBUG] [B2ED6280-367A-4095-95EE-B6D2484234B8] [core] hook=capabilities plugin=cloudron function=hook_capabilities params="" retval=CONT msg=""
Oct 04 09:24:17 [DEBUG] [B2ED6280-367A-4095-95EE-B6D2484234B8] [core] running unrecognized_command hooks
Oct 04 09:24:17 [DEBUG] [B2ED6280-367A-4095-95EE-B6D2484234B8] [core] running unrecognized_command hook in tls plugin

my.xxxxxxx.com
Oct 04 09:24:46 [NOTICE] [B2ED6280-367A-4095-95EE-B6D2484234B8] [tls] timeout setting up TLS
Oct 04 09:24:46 [INFO] [B2ED6280-367A-4095-95EE-B6D2484234B8] [core] hook=unrecognized_command plugin=tls function=upgrade_connection params=STARTTLS retval=DENYSOFTDISCONNECT msg=""

I can't figure out if gmail is causing this or the cloudron mailbox. Any guidance towards the right direction would be so helpful.

Current version is v7.3.0

@girish Thanks for getting back. Yea I was able to sort this out. It appears it might have been related to the route I had established for internet traffic.

Strange that it would only effect gmail, but this new route is simpler, and testing shows all mails being delivered without issue.

This might be a feature but using a volume seems to make my primary mount sda2 think it's completely full.

Main is the name of my volume. Is there a way to remedy this as I can't really install or move apps to sda3 because it thinks sda2 is full.

@girish

I had one app fail to start up because it was on sda2. It could not start because it was out of space according to the readout before it said "not available yet"

I rebooted cloudron and it wouldn't complete startup. So I essentially had to wipe the drive and start over just to get using it again.

On this new rebuild of cloudron it is still doing the possible display bug, but also still appears like it's slowing filling up again as I add my stuff back in. I have lots more to still upload back into the server so I will know soon if it's still a problem.

@nebulon Thanks guys.

@ccfu oooh ok. The list won't become active until it hits a real box. Ok, that makes a lot more sense now. I appreciate it.

@nebulon
I have sent the email titled "Jellyfin test instance for CUDA" all the details for access are in it.

@girish as requested

{"BlockIO":"171GB / 83.4GB","CPUPerc":"0.14%","Container":"2058a37598b0","ID":"2058a37598b013a36a5a3d05a362e96dc4ad212dec1d5db162e372979bc47c38","MemPerc":"84.77%","MemUsage":"24.68GiB / 29.12GiB","Name":"7d771f27-4c9e-48ff-b0ff-3c742305adb3","NetIO":"1.27GB / 410GB","PIDs":"57"}

@nebulon I don't mind at all, I will make some edits to my ACL's rules to allow ssh and send you the details in the email.