Hi, I just upgraded to PhpEd 6.0 Build 6026. I use SFTP with my servers for opening and saving files. In the previous PhpEd version I had no problems with the SFTP functionality, it was very transparent to the user.

But since upgrading to 6.0 Build 6026, I get a lot of errors when opening and saving files: "Socket error # 10053 Software caused connection to abort".

I think this occurs when the SFTP connection times out due to inactivity. But then I have to try refreshing the directory and other stuff several times to get it to reconnect. I think this still occurred in the previous version, but it automatically just reconnected and maybe showed the error in the 'transfers' log, but didn't pop up a message to the user every time to tell them.

I'm finding this pop up and having to retry several times kind of irritating and almost enough to make me revert to the previous version. Can this problem be fixed to make it act more like the previous version? Maybe there is something in the settings somewhere....

Going by other peoples posts, it looks like SFTP might be broken in some respects. I recommend reporting to NuSphere anything that looks like a bug. http://shop.nusphere.com/customer/support.php

However, I've just configured an SFTP account in PhpED 6 build 6026 on a Windows Vista x64 system, going to SSH-2.0-OpenSSH_4.3 on a CentOS 5.5 server and so far it appears to be working perfectly. So when reporting issues to NuSphere it would be a good idea to be specific about your configuration to help them identify the cause.

Thanks plugnplay. I followed your advice and submitted a bug report but I never heard anything back except an automated reply.

I'm using Windows 7 on two different computers, going to the same servers running CentOS 5.

I downloaded build 6028 but it does the same thing. It's very frustrating since it was pretty much flawless in the previous version.

It seems after the SFTP connection times out or gets closed by the server, then it starts popping up those TCP socket errors. Then I have to refresh the directory several times, and usually it complains it can't find the file, then it collapses the entire directory structure (which is very irritating when you are let's say 20 directories deep working on a file), then I have to navigate back to the folder where I was and usually then it works until the next timeout.

How long does it take to time out, and do you know if it does a time out from the Explorer panel or only when uploading/downloading files from within a project?

If it fails from the Explorer panel and I know how long to wait, I can test to see if I also get timeouts or something similar.

@porcupine73

> but I never heard anything back

Normally, it takes up to 1 working day to receive a response from our support.
You may want to check your Junk folder and/or your spam filters and wait a bit if you sent today.

We see that recently google started moving all our mails into junk very easily, not sure why. We never sent any spam.

If nothing helps, please let me know your ticketID and I'll check its status.

Thanks for the fast response. Sorry about that, you're right, I did get a reply I just didn't see it. Ticket ID: WDX-127482. It asked me to try build 6928 (I assume it meant 6028). I am using 6028 right now and it pops up that message.

It normally happens if I haven't saved or opened a file in say 10 or 15 minutes or so. Let me try it right now so I can give you better information.

I'm connecting to the same servers I was using before in the previous version 5 something, and I never saw this message.

I'm have these set up under 'file transfer accounts'. It uses SSH settings to log in with a public key file (no password). These settings I did not change when upgrading to version 6.

So here's the test I'm doing.
1. Created a test.php file and saved it in a directory on the server. (CentOS 5). Time 9:38 AM EST.
2. Waited a while, it's now 9:54AM EST. Went to file save. Got a pop up window
'Error - Failed to check if directory "/apps/" exists
Socket Error # 10053
Software caused connection to abort.
3. Pressed OK
4. It looked like it was saving, but after a couple seconds the same message popped up again
'Error - Failed to check if directory "/apps/" exists
Socket Error # 10053
Software caused connection to abort.
5. Pressed OK
6. Tried to save file again.
7. It looks like it saved it.

So that's just an example for saving a file. I've had this happen when opening a file too, then several times it complains the file doesn't exist, etc, then it collapses the entire directory structure, and after I retry enough times it will finally work.

Now I think these timeouts did happen in the previous version, but I never saw a popup message every time, I think they just showed in the 'transfers' log but it reconnected automatically in the background (and most importantly, silently!)

Please let me know if I may provide any additional information. Thanks.

Thank you for the details, but I'm not sure what to do with them. This forum is not a support channel. Please reply to the ticket with these details.

I'm encountering the same issue. I filed a ticket as well, just so they can start getting an idea of the number of people experiencing the problem.

I sent the above info back with my ticket.

It is very frustrating with those popup messages and then the funky way it acts for several tries after getting that message until it 'reconnects'. Sometimes even after it appears to have reconnected and I"m browsing the folders again, then I double click the file to open, it pops up that message yet again. I never saw that message in the previous version. It's almost frustrating enough to make me revert to the previous version but I'm hopeful this issue will get addressed in the next build.

This might be promising! I did some research and decided to make the following change in /etc/ssh/sshd_config on one of my servers (CentOS 5):



Then I did service sshd restart

Now I did my test, I opened a file in PhpEd, and waited a good 40 minutes. Then I tried to save the file. It saved, and I got NO POPUPS! It did take it about 15 seconds to save, maybe because it was reconnecting or some other reason, but I saw no errors in the transfers log and no popups. So hopefully that will take care of it. I'm forwarding this to my support ticket just for info. Thanks!

FYI: 10053 (connection abort) may mean that your (or your ISP's) router was overloaded wiht # of simultaneous TCP connections and aged off the ssh connection, so it returned abort upon next ssh packet sent to non-existent connection. If it was a ssh daemon on server site, it would normally call socket shutdown and cause 10054 (connection reset) error. The problem with "extra" looping (instead of plain reconnection) in this case will hopefully be fixed in build 6030.

ClientAliveInterval 180 means sending keepalive packets each 180seconds which may help keep connection looking "new" and less preferrable for ageing-off. Anyway it's not reliable.

Thanks for the info. I'm not sure. I'm the only person who has access to ssh / sftp on my servers; they're restricted by Cisco firewall rules. I do sometimes use multiple programs, like I might have one or two WinSCP instances open for file transfers, a Putty SSH session, and then also PhpEd.

Got the message about a test connection. I have a micro instance I sometimes run over at Amazon EC2, I can try to duplicate the problem on that. Opening up a test connection on my production servers would be some amount of effort because I'd have to add exceptions to the Cisco rules, iptables, generate a key (password login is set to not permitted; it's RSA key only), etc. But I can try it on my test server at Amazon EC2 and if it does the same thing there a test account I can then set up easily.