CrushFTP Support Forums
  [Search] Search   [Recent Topics] Recent Topics   [Hottest Topics] Hottest Topics   [Members]  Member Listing   [Groups] Back to home page 
[Register] Register / 
[Login] Login 
Messages posted by: spinkb
Forum Index » Profile for spinkb » Messages posted by spinkb
Author Message
That is a very complex issue, and not something I can help describe. It can be done, but its not something that we are going to encourage or support. If you know how to do it, you would do it...but we unfortunately can't spend the time to teach something as complex as that.

The built in forms will do what you asked having a displayed value which is different from what is submitted. They can also do more complicated cascading features to show different items based on prior from item selections, making the from much more dynamic.

Thanks,
Ben
On 7.1, use Chrome.

Login, view the javascript console. What error does it have?

What are the filename sin this root folder? Special characters?

Thanks,
Ben
What is logged then in the CrushFTP.log when you try HTTPS in the browser?

Are you able to get to the admin interface? If your saying a browser doesn't work, then how did you login as an admin and change your email settings?

Thanks,
Ben
You decide what ports you care about. These are the defaults. I would say most people do:

21,80,443, and 22/2222 depending if opens is using port 22.


pducharme wrote:I can see multiple Ports :

PORT=9090
PORT=21
PORT=2222
PORT=8080
PORT=443

Are all those ports required ?
This just means its your first ever launch and it found no prefs anywhere to load...

pducharme wrote:I also see that in a different docker container (it create 2?) :

java.lang.NullPointerException
at crushftp.server.ServerStatus.SG(ServerStatus.java:4601)
at crushftp.handlers.PreferencesProvider.check_code(PreferencesProvider.java:104)
at crushftp.server.ServerStatus.init_setup(ServerStatus.java:312)
at crushftp.server.ServerStatus.(ServerStatus.java:138)
at CrushFTPD.(CrushFTPD.java:17)
at CrushFTPLauncher.(CrushFTPLauncher.java:87)
at sun.reflect.NativeConstructorAccessorImpl.newInstance0(Native Method)
at sun.reflect.NativeConstructorAccessorImpl.newInstance(NativeConstructorAccessorImpl.java:57)
at sun.reflect.DelegatingConstructorAccessorImpl.newInstance(DelegatingConstructorAccessorImpl.java:45)
at java.lang.reflect.Constructor.newInstance(Constructor.java:526)
at CrushFTP.main(CrushFTP.java:54)
The command runs docker in the background, the -d command. So it will return instantly.

8080 is the default http port. The "-P" param should be allowing you access to that port...but check the "docker ps" to see if the port is being remapped for you by docker, if so, you need to do the custom docker -p 127.0.0.1:5000::5000 mapping style syntax to remap the port.

Did you do the volume mapping correctly? Does your folder now have a CrushFTP.log with messages in it?

Thanks,
Ben
I've thought about your request for quite a while now, staring at it almost daily in the support inbox queue.

I'm not sure it would be a good thing to add something specifically like this in CrushFTP. Its already possible, and maybe just a mini guide on how to get what your looking for is what we should do...but for now, I'll explain it here.

The TempAccount user is the user that applies to all shares that are done. So what you would do is create a new "upload" event on the TempAccount user. Set the FROM to be "%user_emailTo%" and the TO to be "%user_emailFrom%". If you like you can also do a download event too with the same config.

You can also specify a custom upload from be used for TempAccounts in the User Manager, WebInterface section, custom forms.

Now your end user "Bob" can make a folder called "ben", and right click share it, type in Ben's email, enable full access, and send. Ben gets the email, clicks the link, drags a file into the browser, fills out the custom from (if any) and logs out. Bob gets an email that a file has been uploaded by Ben, and he goes and checks it out.

So your "non authenticated" user is someone who clicks the public link that Bob gave them to allows receiving a file.

So you can already do what your asking for, with a ton of customizability to get it just the way you want it too, perform extra actions, etc.

Thanks,
Ben
OK, here is what you need to do.

Download CrushFTP, unzip it. Note where you have this folder at since we will use this when we launch docker. For this example, I am assuming /home/CrushFTP7_PC




So...the whole idea of docker is interesting, but rather pointless in many ways in my opinion. You don't want CrushFTP in your image, or else your data just goes 'poof' when the vm stops...unless you commit it afterwords. So your mounting a local folder into the vm always, and forwarding ports too, always.

In effect, all this docker has that I install is "java". You have CrushFTP, and you just execute java inside the docker. Probably other dockers you could use for this too...

To make your first default admin user....



So you can now use CrushFTP in a docker, but I don't really see the benefit, might as well run java locally, and run CrushFTP natively. If you can run docker, you can likely run java anyway.

Thanks,
Ben
I cannot replicate ethos issue here...there must be something else going on.

I can verify the files are in use while uploading here, but once the upload completes, the files are not in use.

Plus, this would be a very big and widespread problem if this was affecting everyone.

Why do you think they are still in use and locked?

Does the file have a size to it after uploading?

Thanks,
Ben
Try the latest build now and verify this fixes the issue for you.

Click update now on the about tab to get a newer build.

Thanks,
Ben
Set your pref,s logging, debug level to 2, and repeat the test. Let me know what is logged.

Thanks,
Ben
Please give me the full exception message from the CrushFTP log if you can, that way i can find where this occurred at.

Let me know,
Ben
Why don't your try v 7. 1 and see if you still have an issue?
I don't quite understand what your meaning here...

Can you explain this more?
What exact version is this?

What protocol is being used?

Thanks,
Ben
I believe when making the select item, you can separate display and value with a colon and it will be honored when rendering it.

On your html page...I don't get what your asking really, explain more please.

Thanks,
Ben
Put a wait task before the rename...wait for say 3 minutes, then the rename.

Or rename, on failure, wait 3 minutes, another rename task (not looping back to the other one), if that too fails, then email yourself an email.


rename1 -> move
---on fail: wait 3 minutes -> rename2 -> move
-------------------------------------on fail: email admin
Unlikely CrushFTP holding it open...

Virus scanning software?

Can you manually rename it later on?
The rename is failing due to a permissions issue on the OS, or the user has the file still in use...and your OS doesn't allow renaming an in use file (windows).

Its one of those two reasons...or the file didn't exist due tot he user renaming it or something...

Thanks,
Ben
The expiration task cannot wait that long.

Instead use expiration to copy the files to another folder, don't preserve modified date, and have a folder monitor set to cleanup that folder when times reach 30 days old.


How do you know the task hasn't been running? What is in your log?

Thanks,
Ben
Do you have the localization selector?

Can you change to the german localization and does it start working then?

Thanks,
Ben
Expired and enabled are unrelated.

A user can't login if the account is expired, or its disabled.

Thanks,
Ben
And just to followup and confirm your expectation on the HomeDir plugin, i wanted to provide more about the features in it.

You can set the unique time stamp, and prepend the username. If the user were to login twice in a time period that still creates the same home folder, then it will be erased if the allow access to old home folders is not enabled.

The allow access doesn't mean it will let the user access prior date stamped home folders, it means their home folder won't be erased on login if you have it set to always generate.

As far as I can tell looking at the history of this plugin, it has been this way forever. So if you were expecting something else, then that would be wrong.

Thanks,
Ben
This issue may have been a bug with our SFTP VFS. Was this with SFTP? We just fixed a bug with it.

Let me know,
Ben
I think new members aren't allowed to put screenshots is all...over time you will be allowed.

Can you email us directly at support and include the screenshots so we can see the differences?

Are you using the extra vfs, or other linked vfs scenarios?
1.) Delete your user/pass for the email. Then exchange will allow the connection. v5 always tried first with no auth, then tried auth. v6 and v7 just honor your auth settings, and they need to be blank.

2.) Do you have protocol restrictions for the users only allowing one protocol? Or max logins configured on the user?

Thanks,
Ben
The on Connect event can trigger a crush task, and an email event in it, you can then reference each from relents value with the {item_name} type reference, or {this_is_my_field} etc. Everything inside the curly brackets is the form elements name.

Thanks,
Ben
That is all apache config stuff...I'm not an expert in it. I fight with apache every time I have to make changes in my own config...

Now I use CrushFTP as my reverse proxy, and redirect from it to other domains.

But you can do this in apache...just have to get your config right. Order matters, location of config matters (inside a <virtual tag or something like that, I forget. I'm just not familiar with exact syntax in apache...that is outside the scope of CrushFTP, and its something you have to configure yourself. CrushFTP is working fine, its just your apache config is forwarding more than you want to it.

--Ben>
Apache can do different configs based on a domain...that is all in your apache config.

So its possible, you just have an issue in apache currently.
If HTTPS is also owned by Apache, you have to do the exact same scenario for that port as well.

I might suggest just a bit of custom javascript on the login page to look to see if its HTTPS or not, and redirect their browser page to https if its not.

Thanks,
Ben
Update to 7.1, this should fix the issue.

There is something going wrong with your sessions here, and 7.1 has a lot of fixes for sessions.

Thanks,
Ben
In the prefs, http port, advanced...your reverse proxy path should be '/' and not '/crushftp/'.

Thanks,
Ben
If you hard code the CrushFTP IP, you have an issue in that 127.0.0.1 isn't being used so you have to point Apache to the hard coded IP. By leaving it as "lookup", its bound to all IPs and is fine.

When you get the selected resource was not found, we need to see the CrushFTP.log entries, and we need to see *exactly* what you put in your apache config, and the CrushFTP prefs, http port, advanced, reverse proxy config area.

Thanks,
Ben
Still a little confused...

So you a re using homedir, and you have it enabled to allow access to old homers, and its sometimes allowing old access, and sometimes not?
I'm not quite sure what you mean by detecting the home folder...there is no method that does what you describe there.

Please explain more.

Thanks,
Ben
The latest build now supports multithreaded listings on SFTP too. So both FTP and SFTP can do this.

However, in my testing, the OpenSSH client doesn't bother displaying the listing until it has all results...it caches it in memory before showing it. This isn't because of CrushFTP not providing it in chunks though anymore, as we just added that capability, and verified in t eh code it is actually occurring.

So its up to your SFTP client to do multithreading to display the list as its incoming.

7.1.0_18+

Thanks,
Ben
What exact version is this?

Thanks,
Ben
Look at the CrushFTP.log. It should be logging all those failed attempts...so you know what machine tried to connect, and what surname its trying with to give you an idea of what process was expecting to be able to connect.

Thanks,
Ben
Try FTP as a test, does it work there? That will help us to know if the multithreading is working.

The web stands no chance of displaying this much without a massive memory chunk assigned to CrushFTP.

Let me know,
Ben
Change the CrushFTP port 443 to be 4443.

Then try your HTTPS port. If you can still connect...even if it drops you quickly, then something else is running using port 443. I would assume IIS.

Remove IIS, or change IIS to use some other port so CrushFTP can get the port instead.

Thanks,
Ben
That sounds more like your IP has been banned, or you tried configuring the SSL to use a non existent keystore or something besides the defaults.

Out of the box, HTTPS is operations and ready.

To make your server trusted, you definitely need to purchase one, but its functional with your usual browser warnings at any time.

What is logged on your attempts, and what exact CrushFTP version is this?

Thanks,
Ben
Only by having some custom javascript in the prefs, WebInterface, area.

You could make changes there sand change the message you impose using javascript.
I'll look at it, I have not used docker before. First glance it doesn't look like a good fit... But I'll try more.

CrushFTP is already running as a java process. Therefore its OS independent, and mainly just a zip you expand. No real dependancies. I understand the need for what they are doing for some other larger projects that are just a mess of open source tools and such...but CrushFTP isn't that complex in how you set it up and run it. You could literally just have a zip you expand and double click on if you like...

I'll report back here on what I decide.

--Ben
This suggest some and data in your sessions.obj file.

When adjusting the server memory, stop the service, delete sessions.obj, and start the service.

Thanks,
Ben
I suggestion the service/wrapper.conf file, adjust the memory line to be 1024 or 2048.

Possibly there is something causing the server to run out of memory and stop working...then the wrapper restarts CrushFTP when it detects the server is down.

This would explain customers having strange issues too...

Thanks,
Ben
The 127.0.0.1 messages happen for all SFTP connections. They are normal.

These messages I want to see more details when why occur...

java.lang.ClassNotFoundException: crushftp.server.ServerSession
java.io.IOException: Unable to establish loopback connection

What OS is this?

If you look at the admin in CrushFTP, what is the memory usage like, number of threads like?

Thanks,
Ben
I'll consider this for a future update though, we are planning to redo the reporting system.

--Ben
Its unlikely to be possible. Maybe for HTTP(s) it can as that includes some additional headers for the client's true IP. That way a client can be banned correctly, but you always have to have the load balancer in the never ban list to protect things.

The load balancer is using the actual IP...so that is the only IP CrushFTP can ever see.

Thanks,
Ben
There isn't a way to choose where it redirects to...

If your on 7.1_13+ though, it will start its search for the first http port it finds from the current HTTP port its on. So just by organizing the HTTPS port to be after the HTTP port, you will have success.

Thanks,
Ben
Waiting for your direct email to support. We have a solution for you.

Just click update now and you will have a new field to control the MAC algorithm just like the cipher algorithm list.

Thanks,
Ben
 
Forum Index » Profile for spinkb » Messages posted by spinkb
Go to:   
Powered by JForum 2.1.8 © JForum Team