hardy004 has asked for the wisdom of the Perl Monks concerning the following question:

Hi Perl Monks, I have a setup where through my Windows PC(cygwin) I am communicating to the linux system over ssh using Net::Expect::SSH The communication has work really good where i ran 100s of tests. In my tests, after every test failure I reboot the linux system. Intermittently, after reboot, when i try to communicate to the linux system over SSH, I see the below error of "No Pty allocated". I have no clue whats going on ?

2 [main] perl 3528 tty_list::allocate: No pty allocated pty_allocate(nonfatal): posix_openpt(): No such device or address at / +usr/lib/perl5/site_perl/5.10/i686-cygwin/IO/Pty.pm line 24. 2133 [main] perl 3528 tty_list::allocate: No pty allocated pty_allocate(nonfatal): getpt(): No such device or address at /usr/lib +/perl5/site_perl/5.10/i686-cygwin/IO/Pty.pm line 2 4. 4249 [main] perl 3528 tty_list::allocate: No pty allocated pty_allocate(nonfatal): openpty(): No such file or directory at /usr/l +ib/perl5/site_perl/5.10/i686-cygwin/IO/Pty.pm line 24. 6331 [main] perl 3528 tty_list::allocate: No pty allocated pty_allocate(nonfatal): open(/dev/ptmx): No such device or address at +/usr/lib/perl5/site_perl/5.10/i686-cygwin/IO/Pty.p m line 24. Cannot open a pty at /usr/lib/perl5/site_perl/5.10/Net/SSH/Expect.pm l +ine 120

. I realized that there are lot of orphaned SSH connections that i need to kill. Even after doing that, i still am seeing this issue

Could someone please advice ? Thanks.

Replies are listed 'Best First'.
Re: No pty allocated
by eye (Chaplain) on Apr 26, 2014 at 16:16 UTC
    If this is intermittent and only occurring after a reboot, I would guess that you are encountering some sort of race condition. I'm not saying this should happen, but solving it might not be worth too much effort.

    Why are you rebooting the linux host after each failed test? If you really think test failures are doing bad things to the server, you could restart ssh after failed tests. Probably overkill, but should be much easier/faster than rebooting the host.