Hide Forgot
Description of problem: telnet sessions ignore SIGPIPE because systemd defaults to IgnoreSIGPIPE=yes Version-Release number of selected component (if applicable): telnet-server-0.17-59.el7.x86_64 How reproducible: Always Steps to Reproduce: 1. Telnet to a RHEL7 system 2. Run $ perl -le 'print $SIG{PIPE}' 3. Run $ seq 1 100000|head -n 1 Actual results: $ perl -le 'print $SIG{PIPE}' IGNORE $ seq 1 1000000|head -n 1 1 seq: write error: Broken pipe $ Expected results: $ perl -le 'print $SIG{PIPE}' $ seq 1 1000000|head -n 1 1 $ Additional info: /usr/lib/systemd/system/telnet@.service needs a line with IgnoreSIGPIPE=no [Service] IgnoreSIGPIPE=no <======= to be added ExecStart=-/usr/sbin/in.telnetd StandardInput=socket
I was trying to reproduce this on current RHEL-7 and on a few older versions and it works, there is no broken pipe. I had the same versions of telnet, telnet-server and ksh that are described in the customer case. Is this still an issue for you? Because I really cannot reproduce this behaviour.
(In reply to Michal Ruprich from comment #5) > I was trying to reproduce this on current RHEL-7 and on a few older versions > and it works, there is no broken pipe. I had the same versions of telnet, > telnet-server and ksh that are described in the customer case. Is this still > an issue for you? Because I really cannot reproduce this behaviour. Are you sure there wasn't already a "IgnoreSIGPIPE=no" in the telnet unit file already when you tried it? That seems to be the problem. On the other hand, if you put "IgnoreSIGPIPE=yes" in your telnet unit file, are you then able to reproduce the problem? I don't mean to insult you, but you are telnetting into the system, right? If not, that would certainly be the problem, right there. Is there a reason we can't add "IgnoreSIGPIPE=no" in the telnet unit file? Blocking pipe signals is definitely NOT normal behavior for traditional in.telnetd (or xinetd, or inetd, or any of that). Classically, pipe signals are not blocked from a login shell like this, and if someone is using telnet, they are certainly looking for classical behavior, bless their little hearts.
(In reply to Michal Ruprich from comment #5) > I was trying to reproduce this on current RHEL-7 and on a few older versions > and it works, there is no broken pipe. I had the same versions of telnet, > telnet-server and ksh that are described in the customer case. Is this still > an issue for you? Because I really cannot reproduce this behaviour. I just tried a fresh install using rhel7.3 media, telnet-server is now 0.17-60. I can see the exact same behavior from my 1st comment. Note that I use bash (default) and not ksh; the bare $ prompt was just me reducing the info to the minimum. I tried creating an user with ksh as shell, and the results are the same.
> Are you sure there wasn't already a "IgnoreSIGPIPE=no" in the telnet unit > file already when you tried it? That seems to be the problem. I'm sure it wasn't there. > On the other hand, if you put "IgnoreSIGPIPE=yes" in your telnet unit file, > are you then able to reproduce the problem? I even tried that just in case and still could not reproduce. > I don't mean to insult you, but you are telnetting into the system, right? > If not, that would certainly be the problem, right there. Yes I'm using telnet ;) > Is there a reason we can't add "IgnoreSIGPIPE=no" in the telnet unit file? No I don't have a problem with that, I was just trying to confirm the described behaviour. I have a similar bug on rsh where the broken pipe problem actually appears and the IgnoreSIGPIPE fixes it so I think I can fix this one as well.
(In reply to giulioo from comment #7) > I just tried a fresh install using rhel7.3 media, telnet-server is now > 0.17-60. > I can see the exact same behavior from my 1st comment. I did exactly the same but no wrong behaviour. Nevertheless like I said in comment #8 I will fix that, because I see no harm in that.
(In reply to giulioo from comment #7) > [...] > Note that I use bash (default) and not ksh; the bare $ prompt was just me > reducing the info to the minimum. > I tried creating an user with ksh as shell, and the results are the same. Right. This really doesn't have anything to do with the shell. The problem is with telnet. I'm sure csh, tcsh, ash, or any other shell that splits a pipeline into multiple processes would all have similar problems. They may or may not produce an error message, but the underlying problem would still be there. (In reply to Michal Ruprich from comment #9) > [...] > I did exactly the same but no wrong behaviour. Nevertheless like I > said in comment #8 I will fix that, because I see no harm in that. Thank you, Michal. Thanks, tg.
Just a note: I was finally able to reproduce this with clean machine. Apparently my own virtual machine had some older settings that were messing with the reproducer.
Red Hat Enterprise Linux version 7 entered the Maintenance Support 1 Phase in August 2019. In this phase only qualified Critical and Important Security errata advisories (RHSAs) and Urgent Priority Bug Fix errata advisories (RHBAs) may be released as they become available. Other errata advisories may be delivered as appropriate. This bug has been reviewed by Support and Engineering representative and does not meet the inclusion criteria for Maintenance Support 1 Phase. If this issue still exists in newer major version of Red Hat Enterprise Linux, it has been cloned there and work will continue in the cloned bug. For more information about Red Hat Enterprise Linux Lifecycle, please see https://access.redhat.com/support/policy/updates/errata/
Development Management has reviewed and declined this request. You may appeal this decision by using your Red Hat support channels, who will make certain the issue receives the proper prioritization with product and development management. https://www.redhat.com/support/process/production/#howto