How is Net::SSH v2 coming along, any ETA? I am curious to try it out
and see if it fixes my issues with not being able copy files to my
servers with cap.
On 7/26/07, Jamis Buck [EMAIL PROTECTED] wrote:
Interesting stuff, Matt. Note, though, that if the problem is in
OpenSSH, then it won't
Net::SSH and Net::SFTP v2 are coming along very nicely, I'm really
pleased by how they are turning out. As for an ETA: I've got a baby
coming at the beginning of September, and I know if I don't have them
done by then, they won't be done until the end of the year, at the
earliest...so I'm pushing
On 26.7.2007, at 22:17, Jamis Buck wrote:
I'm still suspecting Net::SSH, myself, and it's wild and unnecessary
use of threads. The good news, though, is that Net::SSH v2 is coming
along nicely, and uses IO.select instead of threads, which it should
have done all along.
Jamis,
From one
Well, since the rsync behaviour is known and has been discussed on
cygwin mailing list for years and no alternative was ever given, I
guess there is none in general case.
Note that the specific problem (AFAIK) was that select() can not tell
whether write() would block on a pipe. That is,
I wonder if this SMP/OpenSSH bug could be the cause of the corrupted
mac issue. I get it a lot, as well as Connection reset by peer
errors when doing a deploy via checkout instead of via copy, so I'm
screwed no matter which way I turn :( It should be easy enough (?) to
test this out by running
Yeah, I back up my server onto a old home linux box every hour and I see this
Corrupted Mac on
input problem at least once or twice every day. Everything I could find about
it online pointed to
a hardware issue.
Ben
Edgardo Hames wrote:
Hello guys,
Last year I was working at a
Thanks Jamis for the update, if you need any help testing out patches
or alternative suggestions please let me know - I'd be happy to help
get to the bottom of this one :)
Andy
--
Andrew Beacock
blog: http://blog.andrewbeacock.com
On Jul 19, 2:48 pm, Jamis Buck [EMAIL PROTECTED] wrote:
Andy,
I'd love some suggestions for how to easily duplicate the problem,
actually. I wasn't able to duplicate the problem on that account after
all--it was just uploading really slowly. And it was uploading really
slowly via the OpenSSH sftp utility, too, so it wasn't just the
Net::SFTP or Net::SSH
Andy,
I'm investigating even as we speak. I was set up with an account that
exhibited the problem, and I was finally able to duplicate the problem
there, which has helped. I'll post more as soon as I have more info.
- Jamis
On 7/19/07, Andrew Beacock [EMAIL PROTECTED] wrote:
Jamis,
Have
I'm running 1.8.6. I have the 'fastthread' 1.0 gem for mongrel.
When I did a require for that at the top of my Capfile or my
config/deploy.rb, I still had the client side go into 100% cpu.
I have no problem doing a cap shell to that machine.
Also, I don't think it is a gateway issue. I tried it
On Wednesday 11 July 2007 01:32:58 Ezra Zygmuntowicz wrote:
What version of RUby itself are all of you experiencing this problem
running? Is it perhaps ruby 1.8.6? I've been doing file transfers
with cap here of up to 25Mb without ever seeing this issue no matter
what i call before or
Spoke too soon :(
It worked once (after I added the require line) but not again still
hangs at the same place (** uploading data to ...) and I thought I
was using 1.8.6 but obviously not!
I'm using:
Ubuntu 6.10 (lsb_release -r - 6.10)
Kernel 2.6.17 (uname -a - Linux 2.6.17-11-386 #2 Fri May
Just a little more information:
If I comment out my :web and :db targets (leaving just :app) then 9
out of 10 times the deploy works correctly.
As soon as I add my :db target back in (a different hostname) it hangs
in the same place as before.
Andy.
--
Andrew Beacock
Blog:
Note that 100% CPU consumption is a known issue with Net::SFTP, and is
not necessarily related to the upload hanging. Even when everything
works correctly, you will still see huge CPU usage by Net::SFTP.
- Jamis
On 7/11/07, Andrew Beacock [EMAIL PROTECTED] wrote:
Just a little more
jamis, if it would help.
I could see if my hosting provider can get you access to an instance on
the machine.
On 7/11/07, Jamis Buck [EMAIL PROTECTED] wrote:
Note that 100% CPU consumption is a known issue with Net::SFTP, and is
not necessarily related to the upload hanging. Even when
That would help immensely. As it is, the only error I can duplicate
consistently is a corrupted mac error when performing two or more
simultaneous large uploads via a gateway...which is not the same error
others are seeing.
- Jamis
On 7/11/07, [EMAIL PROTECTED] [EMAIL PROTECTED] wrote:
jamis,
I'm sending you the info in private mail
Cheers
On 7/11/07, Jamis Buck [EMAIL PROTECTED] wrote:
That would help immensely. As it is, the only error I can duplicate
consistently is a corrupted mac error when performing two or more
simultaneous large uploads via a gateway...which is not the
What version of RUby itself are all of you experiencing this problem
running? Is it perhaps ruby 1.8.6? I've been doing file transfers
with cap here of up to 25Mb without ever seeing this issue no matter
what i call before or after the put command. THis was with ruby 1.8.5
Thanks for all the data, that's great stuff. I'm still at a loss as to
what the difference is... especially if it works in cap1 but not cap2,
against the same Net::SSH version. If it isn't too much to ask, could
you try comparing the sftp dumps against a cap1 session, to see if
there is any
will do. i just had to punt to get the app shipped in the meantime.
On 7/9/07, Jamis Buck [EMAIL PROTECTED] wrote:
Thanks for all the data, that's great stuff. I'm still at a loss as to
what the difference is... especially if it works in cap1 but not cap2,
against the same Net::SSH version.
I also have upload problem from WinXp with the latest net-sftp and net-
ssh (1.1.2)
The upload command succeed without error but the .zip file copied to
the server is only 13685 bytes out of 120k .zip file. So the
decompress step fails.
Same problem without gateway or with gateway (the reason I
On Saturday 07 July 2007 23:43:46 Jamis Buck wrote:
So, I suppose my question is: for those of you having problems, what
is your connection configuration like? Are you using a gateway? Are
you connecting to multiple hosts?
No gateway and no multiple hosts.
Here is the configuration i use:
More data, I recompiled the sftp-server with some extra logging thrown in
to see what was happening with cap.
Jul 8 15:22:31 m0 sftp-server[19679]: debug1: request 0: sent handle handle
0
Jul 8 15:22:31 m0 sftp-server[19679]: debug1: process message len 0
Jul 8 15:22:31 m0 sftp-server[19679]:
I am running capistrano on my local box to do the check and trying to
copy the file directly the box I want it without the use of a gateway.
I've gotten Net::SFTP to work just fine using put_file, is there any
special reason cap is read the file in and then using put instead of
using put_file?
The original use-case for put in Capistrano was for writing generated
templates to the remote servers. Thus, the put method takes a string
and writes it. I'd not be opposed to extending cap's put method to
handle file-to-file transfers, though.
- Jamis
On 7/7/07, Joshua S [EMAIL PROTECTED]
I used the foo task above to try and upload the tar.gz created by copy/
export and then run it through the debugger. It seems to be caught in
a loop, I am including a trace of the code that is looping but do not
currently have time to go through it all.
Hi Jamis.
I've tried run method. It makes no difference.
Is there any way to get some kind of traces or other debug info?
So we can see what's the difference between the two cases.
Dmitry
On 19 июн, 01:52, Jamis Buck [EMAIL PROTECTED] wrote:
It shouldn't make any difference at all, but what
Dmitry,
Can you tell me what OS you are running locally? Also, what Ruby
version do you have installed locally? What versions of Net::SSH and
Net::SFTP do you have installed? And how large is mapscript.so?
- Jamis
On Jun 18, 2007, at 12:46 AM, Dmitry Kurochkin wrote:
Hi Capistrano devs.
Hi Jamis,
I'm running Debian unstable,
ruby 1.8.6 (2007-03-13 patchlevel 0) [i486-linux],
net-sftp (1.1.0),
net-ssh (1.1.1).
mapscript.so is 1.4 MB.
Regards,
Dmitry
On 18 июн, 16:59, Jamis Buck [EMAIL PROTECTED] wrote:
Dmitry,
Can you tell me what OS you are running locally? Also, what
It shouldn't make any difference at all, but what if you use run
instead of invoke_command? The invoke_command method is intended
mostly for internal use, or for situations where you need to
programmatically switch between run/sudo.
- Jamis
On Jun 18, 2007, at 3:44 PM, Dmitry Kurochkin
Also, how long have you waited, when it appears hung? Net::SFTP is,
sadly, very inefficient, and will take longer than it ought to for
large files (and 1.4MB is large, for Net::SFTP). Also, it will cause
your CPU usage to spike, which could look like an infinite loop.
- Jamis
On Jun 18,
Some more info.
I have update net-ssh to 1.1.3 and the problem is still there.
Besides all results above are with capistrano 1.99.1.
I have just tried 1.99.2 and put does not work at all. It hangs even
without invoke_command before put.
Thanks,
Dmitry
32 matches
Mail list logo
