Networking :: Pure-ftpd Passive Transfers Not Working?
Jun 4, 2010
I have a dns-323 linux device that's running pure-ftpd with SSL/TLS authentication. Pure-ftpd is sitting behind a linksys router with IP 192.168.1.51. Pure-ftpd is configured for port 8021 and passive port range 55562-55663. The linksys router is configure to forward port 8021 and the passive port range to 192.168.1.51.
From outside my network I can connect to the ftp server using the WAN address of the router. I'm using filezilla 2.2.32 as my client and I choose FTP w/ explicit TLS (no other option will connect). The client will authenticate successfully with pure-ftpd but once it sets up the passive data connection and tries to do a LIST of the root directory, there's a timeout. I'm assuming this is because the passive data connection is not working. In pure-ftpd, I tried changing the passive address that it reports, to be the WAN address of the router, but it did not make a difference. I included the log from filezilla below.
This create a new error. 421 Unable to read the indexed puredb file (or old format detected) - Try pure-pw mkdb I am getting a feeling this is getting from bad to worse. I attempted to remove pure-ftpd and pureadmin to start over again. However, it appears I cannot remove everything. I attempted the following:
apt-get remove pure-ftpd pureadmin I am not sure what this really does because I can still see the folder /etc/pure-ftpd and I cannot remove it. I could not use rm /etc/pure-ftpd because the system said it was not empty. I also tried sudo aptitude remove pure-ftpd pureadmin . Again I cannot completely remove the folder. How do I completely remove pure-ftpd and pure admin and all databases, files and folders so I have a fresh system ready to start the install again?
I have set up pure-authd and pure-ftpd. They are both running, I have created the socket etc.
In my authentication module (a php script) for testing purposes I have done a vardump in to a file, and have realised that pure-authd is not passing on any variables (username, password of the current person trying to log in via ftp) to the PHP script.
I am sure the authentication module is working (have tested it vigorously on the command line), but after 10 hours wondering why it wouldnt work and messing about with the script, I have realised that the variables were never even getting in to the script in the first place!
I am running the processes such as this:
Everything seems as if it is working other than this. For instance, when testing the setup with a very basic auth module which doesnt require a username or password (the basic module just passes "auth_ok:1" to pure-ftpd and the user is then logged in), I can log in to the FTP server fine.
But like I say, a vardump ($argv) on my proper PHP authentication script would suggest that no username or password are being passed to it.
I have a "friend" who hosts a couple of websites that I maintain. I have always used filezilla client to connect and update the files on the website.I lost access for a little while and he said that "The profiles had accidentally been dropped".He reinstated everything but now I cannot complete a log on session. I maintain several other websites which have the same PURE FTPd server at their end and I can connect to these OK.The delicacy of the matter is that, with no changes at my end, and my other sites working, logic suggests that something has changed at his end. I have very patiently worked through all his suggestions but to no avail.I can't suggest that he is at fault when he claims expertise in the field and I am a complete lamer when it comes to FTP.
I wonder if anyone would be kind enough to review the log on result below and see if they can throw some light on the situation for me.(I have anonymised the IP address and other information for security reasons)
I'm really confused with this one. Yesterday our FTP cluster went down, i havent yet tracked the reason but i suspect a security update killed it as automatic updates are enabled (i will be turning those off now!).
Anyway, FTP connections were receiving a 425 Sorry, Unable to resovle error message, i sound the solution to be to disable dns lookups with the DontResolve > yes option.
but for some unknown reason this is only working on my failover box! when i restart pure-ftpd on the primary box the command switch -H is missing?
i've done the same thing on both boxes by creating a file in /etc/pure-ftpd/conf called DontResolve and editing it to just read the word 'yes', been over it and checked for spelling mistakes and i cant find any reason why its ignoring the option?
I am running an FTP server using Pure-FTPD which was a migration from an OS-X workstation to Fedora 10 - VMware esxi.
This was done for various reasons but the reason for choosing the Pure-FTPD app was due to an easier migration.
The system has been migrated for about a month and works very well. I installed it on the server using yum install pure-ftpd, configured it and imported all the accounts [about 150 accounts] so this is all fab. however we now want to see if we can get an upload notification.
This has lead me to look at the uploadscript option in pure-ftpd. we use virtual users and the pure-pw file.
When we set the directive
The ftp server hangs, I cant seem to see how to get this to work, from my investigation I think the issue is that the yum install is not building the app " --with-uploadscript " i have tried to do a hand install on a test Fedora 10 VM, I downloaded pure-ftpd-1.0.27.tar.gz and unpacked it, did a ./configure --with uploadscript then a make and make install, nothing reports any errors but it doesnt seem to install anything. I can start the server from a command line but there are no config directories ? and even after creating a new user and a passwd-mkdb i cant log in to the server. but that most likely as i dont have a full configuration. but i can t find where to put the configuration files.
I'm trying to install Pure-FTPd. I'm going to use it via the internal file-based userlist.The problem is, when I yum install pure-ftpd from rpmforge or CentOS.Karan.Org, I also get mysql, perl-DBI and postgresql-libs as dependencies.Since I'm not intrested in DB-based authentication, I'd strongly prefer to skip these packages.Why are these listed as dependencies, since AFAIK, they are not really dependencies? Can I force yum to skip them somehow?
I am having difficulties with Pure-FTPD. I had it working at one time. I recently had to reinstall Ubuntu 10.10 because pure-ftpd stopped working for some reason and now I have a fresh install but pure-ftpd still does not work. This is what I did to install it.
Sudo apt-get install pure-ftpd
Using the Ubuntu Spftware Center I installed PureAdmin. I then create a virtual user. Then in the terminal entered the following:
I read somewhere that this could be a filezilla issue and to fix it I need to use active mode and use the filezilla external address [URL]... Of course I did this and the filezilla solution did not work this time. I am fairly sure it is a server issue this time.
I want to make an Ubuntu box a completely passive sniffer. As such, I want the NIC to never send anything or replying to anyone on the network, while still being able to capture the network traffic using pcap.
Can anybody have an idea of an ftp client code using raw ftp commands.
OR Can anybody tell me how to turn off passive mode using raw ftp commands. PASV is the command which turns on passive mode but there is no such alternative of active mode. I have already tried using pasv_enable=NO in vsftpd.conf but that is not working.
A little over a year ago I was using SCP to successfully transfer large files over my LAN (exact same hardware). I can't seem to do this any more, and I'm not sure why.I think it's either something with iptables, or a network card driver problem. I use the same driver for both computers (b43 wireless). I can't do FTP transfers either. They start going, but quickly stall. I've used the Firestarter (iptables gui) to allow all the correct connections.One last thing: When I tried to connect to ssh using an alfa wireless card (not sure of the drivers), I couldn't even connect to ssh period. Same settings were used.
I have a server set up as an NFS share, and the share mounted on my laptop. Using linksys wireless g router and 15mb internet connection. Laptop is on wireless connection and server is wired.While transferring files from the laptop to the server I only get about 55kb/s. Is this normal for wireless g?
I am running Samba on a debian Lenny box on a wireless home network. I find that file transfers to the samba share are very slow. It takes over a minute to copy a 40MB file to the linux box, but only 20 seconds to copy the same file to a windows XP box on the same network.
Anyways, I could use a little direction on how to proceed with this, I'm really not sure where to start,
I am having a problem with slow data transfers with both Samba and scp. I have gigabit NIC's on both all three machines that I am transferring to and from, connected to a gigabit switch. My data transfers under both smb and scp average around 21 MBit/s, (I am using nload to monitor transfer speeds).The machines are configured as follows,1) desktop
AMD Athlon 64 X2 6000+ 6 gig Corsair memory Realtek RTL8168C(P) gigabit NIC (on board)
I need to transfer 330G of data from a hard drive in my workstation to my NAS device. The entire network is gigabit and being run with new HP procurve switches. All machines have static IP addresses. The NAS is a Buffalo Terastation PRO which has the latest firmware, is set to jumbo frames, and has just been upgraded with 4 brand new 500G drives giving us a 1.4TB raid 5 setup. My workstation is a dual Quad core xeon box running on an Intel S5000XVN board with 8G of ram. My OS is Ubuntu 10.04 x64 running on a pair of Intel X25 SSDs in a raid mirror. The data drive is a 500G SATA drive connected to my onboard controller. The file system on the SATA drive is XFS. This problem was ongoing before I got my new workstation, before we had the GB switches, and before the NAS got new drives. When I transfer a small file or folder (less than 500M) it reaches speeds of 10-11 MB/sec. When I transfer a file or folder larger than that the speed slows to a crawl (less than 2MB/sec). It has always been this way with this NAS. Changing to jumbo frames speeds up the small transfers but makes little difference in the big ones. I verified with HP that the switches are jumbo frame capable.
I tried using ssh between my netbook and desktop, but it was going to take around 30 hours to transfer 39GB over the home network. Also SSH is very sketchy and often drops connections.I've been messing with it all day and I'm quite frustrated.What I'm looking to do is use my netbook as more of a primary computer and the desktop as a storage computer. Not quite a server, because I'd like to still keep a GUI on it. I'd like to be able to keep my music and movies on the desktop and stream them to the netbook (SSH sucks for this, always drops connections). I've already set up the web client for Transmission bit torrent client so I can torrent on a machine that's almost always on and connected.
Is there a better setup for all of this? I like the netbook because of the portability; I like the desktop because it's always connected (for torrents) and it has a larger storage capacity. It would be mainly used around the house. I would like to back up a file or two while abroad, but I'm not looking to stream music while I'm across town or anything.
I have a tenda wireless adaptor running on 11.04 natty. The USB id is: 148F 3070
Using the already installed drivers rt2800usb it can connect to all 13 channels but is very flaky, transfers slowing and stopping regularly. The rt2870sta driver is a lot more stable, but it will only connect to channels 1-11, but I need to use channel 13 due to massive wifi congestion on other channels...
I've tried iwpriv but it says there are no ioctls for the device.
Is there any way to get the installed driver rt2870sta to scan and connect to channel 13?
I've also tried to install the latest drivers from the ralink website: rt2870sta says it can see all 14 channels, but fails to scan. I also tried rt3070 driver but I cannot insmod as there are errors...
I just installed ubuntu as my primary OS, but I have the disk with XP on it and I don't want to go back, but I need faster network connectivity. I have a T60p with Intel Gigabit jacked into my Gigabit router which also has my desktop (running XP) and my NAS. If I FTP files from my NAS (or SCP), I get transfer speeds around 250-500 KB/s (which is not very fast). On this same switch, from my XP desktop I get transfer speeds around 12 MB/s. I get the same speeds using my 802.11n card (Atheros) as with the ethernet NIC (250-500 KB/s).The drivers for the ethernet card and the atheros card are e1000e and ath9k respectively.I have disabled IPv6. Since the problem occurs using either interface, I am just going to concentrate on fixing it for the Ethernet interface (since I believe it to be a systemwide problem).
skinnersbane@albert:~$ sudo ethtool eth0 Settings for eth0: Supported ports: [ TP ] Supported link modes: 10baseT/Half 10baseT/Full
Clearly my card is running at Gigabit, but why the bad transfer speeds? I am using filezilla for FTP (technically FTPES). I closed every other program. My CPU utilization does seem high and I wonder if this is part of the problem. I had no problems with throughput using either interface in Windows XP just one week ago.
after successfully configuring the dwa-552 to work in master mode in ubuntu 10.04 (ath9k driver) I ran some file transfer tests. The download speed is very good (~50mbps) but the upload speed spikes at about 10-20mbps for the first few KB and then it's nonexistent (0-1kbps). This only affects file transfers or otherwise bandwidth consuming processes. Normal web browsing or ssh is not affected. After running a speedtest of my internet connection which is routed through the AP I could upload to the internet with 1mbps which is my inet connection maximum so apparently this is not affected. Tried the same file transfers with netcat to eliminate any other factors and had the same problem. dmesg and hostapd debug did not report anything unusual
I have a Pure Evoke Flow radio that connect to my wireless router for streaming internet radio, this works fine. It also has the functionality to connect to a hard drive over wifi and play any MP3 files on it and it runs on Linux.
With Windows (sorry) I just click on the folder and make it Shareable, on the Radio I then set it to search for servers and it finds my wife's PC and plays her music.
I have (I believe) set my music directory on Ubuntu 9.10 to be shareable, it downloaded and installed a couple of files including samba and then rebooted. However my Radio cannot see it.
my proftpd setup will not do passive ports, reading the how-to on this page [URL] It gives the instructions on how to set up passive ports and masqurade but the example is using ipchains so before I start typing these commands in I would like to know if I need to change anything in the commands other then the ipchains command.
will this work on debian 5.05 also how do I find out if I use ip filters.