George Nachman / iterm2

I see that the servers survive a log-out/log-in but I don't get the brokenness you describe. I tried switching accounts, but that didn't do it either. Something else is going on here; maybe some more detail on what you're doing, or maybe some system preference.

I just experienced this issue myself. I think what happened was:

I logged into a remote server within iTerm2Beta3
I locked my system
when I came back, for some reason, my user was logged out
I logged in
Sessions where restored (I assume; this was a few days ago)
I just logged out of the remote SSH server and tried to log in again (vagrant ssh in my case)
got No user exists for uid 501

I had to close the respective iTerm2Beta3 window, create a new one, then I could log in again.

Note: between step 3/4 and 5/6 there were "long" time gaps, as in, literally a day or so (I never shutdown my system, just close the lid/lock it manually).

Following https://iterm2.com/restoration.html , here's the relevant /var/log/system.log output:

Feb 25 08:08:18 lv426 iTerm2-Server[58882]: select returned -1, error = Interrupted system call
Feb 25 08:08:18 lv426 iTerm2-Server[58882]: Calling select...
Feb 25 08:08:18 lv426 iTerm2-Server[58882]: select returned 1, error = Interrupted system call
Feb 25 08:08:18 lv426 iTerm2-Server[58882]: select returned. child dead=2, connection closed=0
Feb 25 08:08:18 lv426 iTerm2-Server[58882]: Connection closed.
Feb 25 08:08:18 lv426 iTerm2-Server[58882]: Unlink /var/tmp/iTerm2.socket.58882
Feb 25 08:08:18 lv426 iTerm2[78292]: File descriptor server exited with status 0
Feb 25 08:08:19 lv426 iTerm2[78292]: ReceiveMessageAndFileDescriptor
Feb 25 08:08:19 lv426 iTerm2-Server[34068]: Installing SIGHUP handler.
Feb 25 08:08:19 lv426 iTerm2-Server[34068]: Installing SIGCHLD handler.
Feb 25 08:08:19 lv426 iTerm2-Server[34068]: Sending file descriptor and waiting on initial connection
Feb 25 08:08:19 lv426 iTerm2-Server[34068]: send master fd and child pid 34069
Feb 25 08:08:19 lv426 iTerm2-Server[34068]: All done. Waiting for client to disconnect or child to die.
Feb 25 08:08:19 lv426 iTerm2-Server[34068]: Calling select...
Feb 25 08:08:19 lv426 iTerm2[78292]: recvmsg returned 4
Feb 25 08:08:19 lv426 iTerm2[78292]: Got a fd
Feb 25 08:08:19 lv426 iTerm2[78292]: Return 4
Feb 25 08:08:23 lv426 iTerm2[78292]: File descriptor server exited with status 0
Feb 25 08:08:48 lv426 iTerm2[78292]: ReceiveMessageAndFileDescriptor
Feb 25 08:08:48 lv426 iTerm2-Server[34105]: Installing SIGHUP handler.
Feb 25 08:08:48 lv426 iTerm2-Server[34105]: Installing SIGCHLD handler.
Feb 25 08:08:48 lv426 iTerm2-Server[34105]: Sending file descriptor and waiting on initial connection
Feb 25 08:08:48 lv426 iTerm2-Server[34105]: send master fd and child pid 34106
Feb 25 08:08:48 lv426 iTerm2-Server[34105]: All done. Waiting for client to disconnect or child to die.
Feb 25 08:08:48 lv426 iTerm2-Server[34105]: Calling select...
Feb 25 08:08:48 lv426 iTerm2[78292]: recvmsg returned 4
Feb 25 08:08:48 lv426 iTerm2[78292]: Got a fd
Feb 25 08:08:48 lv426 iTerm2[78292]: Return 4
Feb 25 08:08:48 lv426 iTerm2[78292]: Time to encode state for window <PseudoTerminal: 0x7f834eb78650 tabs=1 window=<PTYWindow: 0x7f834ef95f70 frame=NSRect: {{2, 13}, {1916, 1036}} title=4. bash alpha=1.000000 isMain=0 isKey=0 isVisible=1 delegate=0x7f834eb78650>>: 0.0003660321235656738
Feb 25 08:08:48 lv426 iTerm2[78292]: Time to encode state for window <PseudoTerminal: 0x7f834e8f36b0 tabs=1 window=<PTYWindow: 0x7f834bfaf650 frame=NSRect: {{2, 13}, {1916, 1036}} title=2. vagrant@dev: ~ (vagrant) alpha=1.000000 isMain=0 isKey=0 isVisible=1 delegate=0x7f834e8f36b0>>: 0.0008410215377807617
Feb 25 08:08:48 lv426 iTerm2[78292]: Time to encode state for window <PseudoTerminal: 0x7f834bdc0450 tabs=1 window=<PTYWindow: 0x7f834bc6e4d0 frame=NSRect: {{641, 12}, {836, 1036}} title= alpha=1.000000 isMain=1 isKey=1 isVisible=1 delegate=0x7f834bdc0450>>: 0.0003739595413208008
Feb 25 08:08:54 lv426 iTerm2-Server[34105]: select returned -1, error = Interrupted system call
Feb 25 08:08:54 lv426 iTerm2-Server[34105]: Calling select...
Feb 25 08:08:54 lv426 iTerm2-Server[34105]: select returned 1, error = Interrupted system call
Feb 25 08:08:54 lv426 iTerm2-Server[34105]: select returned. child dead=2, connection closed=0
Feb 25 08:08:54 lv426 iTerm2-Server[34105]: Connection closed.
Feb 25 08:08:54 lv426 iTerm2-Server[34105]: Unlink /var/tmp/iTerm2.socket.34105
Feb 25 08:08:54 lv426 iTerm2[78292]: File descriptor server exited with status 0
Feb 25 08:08:56 lv426 iTerm2[78292]: Time to encode state for window <PseudoTerminal: 0x7f834e8f36b0 tabs=1 window=<PTYWindow: 0x7f834bfaf650 frame=NSRect: {{2, 13}, {1916, 1036}} title=2. vagrant@dev: ~ (vagrant) alpha=1.000000 isMain=0 isKey=0 isVisible=1 delegate=0x7f834e8f36b0>>: 0.0005910396575927734
Feb 25 08:08:59 lv426 iTerm2[78292]: File descriptor server exited with status 0
Feb 25 08:09:27 lv426 iTerm2[78292]: Time to encode state for window <PseudoTerminal: 0x7f834e8f36b0 tabs=1 window=<PTYWindow: 0x7f834bfaf650 frame=NSRect: {{2, 13}, {1916, 1036}} title=2. vagrant@dev: ~ (vagrant) alpha=1.000000 isMain=0 isKey=0 isVisible=1 delegate=0x7f834e8f36b0>>: 0.0004370212554931641
Feb 25 08:11:22 lv426 iTerm2[78292]: Time to encode state for window <PseudoTerminal: 0x7f834e8f36b0 tabs=1 window=<PTYWindow: 0x7f834bfaf650 frame=NSRect: {{2, 13}, {1916, 1036}} title=2. vagrant@dev: ~ (vagrant) alpha=1.000000 isMain=0 isKey=0 isVisible=1 delegate=0x7f834e8f36b0>>: 0.0006260275840759277
Feb 25 08:12:04 lv426 iTerm2[78292]: Time to encode state for window <PseudoTerminal: 0x7f834e8f36b0 tabs=1 window=<PTYWindow: 0x7f834bfaf650 frame=NSRect: {{2, 13}, {1916, 1036}} title=2. vagrant@dev: ~ (vagrant) alpha=1.000000 isMain=0 isKey=0 isVisible=1 delegate=0x7f834e8f36b0>>: 0.0001580119132995605
Feb 25 08:12:07 lv426 iTerm2-Server[34068]: select returned -1, error = Interrupted system call
Feb 25 08:12:07 lv426 iTerm2-Server[34068]: Calling select...
Feb 25 08:12:07 lv426 iTerm2-Server[34068]: select returned 1, error = Interrupted system call
Feb 25 08:12:07 lv426 iTerm2-Server[34068]: select returned. child dead=2, connection closed=0
Feb 25 08:12:07 lv426 iTerm2-Server[34068]: Connection closed.
Feb 25 08:12:07 lv426 iTerm2-Server[34068]: Unlink /var/tmp/iTerm2.socket.34068
Feb 25 08:12:07 lv426 iTerm2[78292]: File descriptor server exited with status 0
Feb 25 08:12:11 lv426 iTerm2[34132]: ReceiveMessageAndFileDescriptor
Feb 25 08:12:11 lv426 iTerm2-Server[34136]: Installing SIGHUP handler.
Feb 25 08:12:11 lv426 iTerm2-Server[34136]: Installing SIGCHLD handler.
Feb 25 08:12:11 lv426 iTerm2-Server[34136]: Sending file descriptor and waiting on initial connection
Feb 25 08:12:11 lv426 iTerm2-Server[34136]: send master fd and child pid 34137
Feb 25 08:12:11 lv426 iTerm2[34132]: recvmsg returned 4
Feb 25 08:12:11 lv426 iTerm2-Server[34136]: All done. Waiting for client to disconnect or child to die.
Feb 25 08:12:11 lv426 iTerm2-Server[34136]: Calling select...
Feb 25 08:12:11 lv426 iTerm2[34132]: Got a fd
Feb 25 08:12:11 lv426 iTerm2[34132]: Return 4
Feb 25 08:12:20 lv426 iTerm2[34132]: Time to encode state for window <PseudoTerminal: 0x7febeb8191b0 tabs=1 window=<PTYWindow: 0x7febe974a350 frame=NSRect: {{2, 13}, {1916, 1036}} title=1. vagrant alpha=1.000000 isMain=0 isKey=0 isVisible=1 delegate=0x7febeb8191b0>>: 0.0001940131187438965
Feb 25 08:12:20 lv426 iTerm2[34132]: Time to save app restorable state: 5.704164505004883e-05
Feb 25 08:12:36 lv426 iTerm2[34132]: Time to encode state for window <PseudoTerminal: 0x7febeb8191b0 tabs=1 window=<PTYWindow: 0x7febe974a350 frame=NSRect: {{2, 13}, {1916, 1036}} title=1. vagrant@dev: ~ (vagrant) alpha=1.000000 isMain=0 isKey=0 isVisible=1 delegate=0x7febeb8191b0>>: 0.0002279877662658691
Feb 25 08:12:36 lv426 iTerm2[34132]: Time to save app restorable state: 6.598234176635742e-05
Feb 25 08:13:19 lv426 iTerm2[34132]: ReceiveMessageAndFileDescriptor
Feb 25 08:13:19 lv426 iTerm2-Server[34178]: Installing SIGHUP handler.
Feb 25 08:13:19 lv426 iTerm2-Server[34178]: Installing SIGCHLD handler.
Feb 25 08:13:19 lv426 iTerm2-Server[34178]: Sending file descriptor and waiting on initial connection
Feb 25 08:13:19 lv426 iTerm2-Server[34178]: send master fd and child pid 34179
Feb 25 08:13:19 lv426 iTerm2[34132]: recvmsg returned 4
Feb 25 08:13:19 lv426 iTerm2[34132]: Got a fd
Feb 25 08:13:19 lv426 iTerm2-Server[34178]: All done. Waiting for client to disconnect or child to die.
Feb 25 08:13:19 lv426 iTerm2[34132]: Return 4
Feb 25 08:13:19 lv426 iTerm2-Server[34178]: Calling select...
Feb 25 08:13:25 lv426 iTerm2[34132]: Time to encode state for window <PseudoTerminal: 0x7febeb8191b0 tabs=1 window=<PTYWindow: 0x7febe974a350 frame=NSRect: {{2, 13}, {1916, 1036}} title=1. vagrant@dev: ~ (vagrant) alpha=1.000000 isMain=0 isKey=0 isVisible=1 delegate=0x7febeb8191b0>>: 0.0002139806747436523
Feb 25 08:13:25 lv426 iTerm2[34132]: Time to encode state for window <PseudoTerminal: 0x7febee2608d0 tabs=1 window=<PTYWindow: 0x7febe94fb780 frame=NSRect: {{2, 12}, {836, 1036}} title=2. mfn@home: ~ (ssh) alpha=1.000000 isMain=0 isKey=0 isVisible=0 delegate=0x7febee2608d0>>: 0.0001339912414550781
Feb 25 08:17:47 lv426 iTerm2[34132]: ReceiveMessageAndFileDescriptor
Feb 25 08:17:47 lv426 iTerm2-Server[34200]: Installing SIGHUP handler.
Feb 25 08:17:47 lv426 iTerm2-Server[34200]: Installing SIGCHLD handler.
Feb 25 08:17:47 lv426 iTerm2-Server[34200]: Sending file descriptor and waiting on initial connection
Feb 25 08:17:47 lv426 iTerm2-Server[34200]: send master fd and child pid 34201
Feb 25 08:17:47 lv426 iTerm2-Server[34200]: All done. Waiting for client to disconnect or child to die.
Feb 25 08:17:47 lv426 iTerm2-Server[34200]: Calling select...
Feb 25 08:17:47 lv426 iTerm2[34132]: recvmsg returned 4
Feb 25 08:17:47 lv426 iTerm2[34132]: Got a fd
Feb 25 08:17:47 lv426 iTerm2[34132]: Return 4
Feb 25 08:17:59 lv426 iTerm2[34132]: Time to encode state for window <PseudoTerminal: 0x7febeb8191b0 tabs=1 window=<PTYWindow: 0x7febe974a350 frame=NSRect: {{2, 13}, {1916, 1036}} title=1. vagrant@dev: ~ (vagrant) alpha=1.000000 isMain=0 isKey=0 isVisible=1 delegate=0x7febeb8191b0>>: 0.0002179741859436035
Feb 25 08:17:59 lv426 iTerm2[34132]: Time to encode state for window <PseudoTerminal: 0x7febee25b2d0 tabs=1 window=<PTYWindow: 0x7febe97e8240 frame=NSRect: {{2, 13}, {1916, 1036}} title=3. bash alpha=1.000000 isMain=0 isKey=0 isVisible=1 delegate=0x7febee25b2d0>>: 0.000129997730255127

Here's the ps output:

~ $ ps -aef|grep iTerm2
  501 34132     1   0  8:12AM ??         0:02.51 /Users/myuser/Applications/iTerm2Beta3.app/Contents/MacOS/iTerm2
  501 34136 34132   0  8:12AM ttys000    0:00.02 /Users/myuser/Applications/iTerm2Beta3.app/Contents/MacOS/iTerm2 --server /usr/local/bin/bash
  501 34178 34132   0  8:13AM ttys001    0:00.02 /Users/myuser/Applications/iTerm2Beta3.app/Contents/MacOS/iTerm2 --server /usr/local/bin/bash
  501 34200 34132   0  8:17AM ttys002    0:00.02 /Users/myuser/Applications/iTerm2Beta3.app/Contents/MacOS/iTerm2 --server /usr/local/bin/bash
  501 34241 34132   0  8:18AM ttys003    0:00.02 /Users/myuser/Applications/iTerm2Beta3.app/Contents/MacOS/iTerm2 --server /usr/local/bin/bash
  501 34261 34242   0  8:18AM ttys003    0:00.00 grep iTerm2

@mfn Can you reproduce this consistently?

I doubt it. As I said, system was suspended/a few days later, etc. The whole session thing come so surprising that I immediately disabled it, sorry.

I can reproduce it. I just logged out and logged back in and I got the same issue. I then typed logout at each terminal and restarted it, and logged out and back in again, and it reproduced again. So it doesn't seem that the length of time that the session is running matters.

Status changed to closed

Commit f300ec3 kills jobs before logging out. Seems reasonable to do, anyway, even if not everybody experiences this issue.

This just happened to me in Build 2.9.20160113. I unplugged my machine at work, shut the lid, brought it home and saw No user exists for uid 501 while trying to reconnect.

It is possible that this is a general mac issue. I have never experienced it with my home machine that is not screen-locked (and I have moved it around a lot). My work machine is screen-locked, so the problem we see in Iterm may have something to do with resuming the GUI session on the mac. In the same instance, I have actually had a screen lock forced on me a couple minutes after I had gone online and started working. After unlocking the screen, I could no longer reconnect my Iterm sessions.

The actual meaning of the message is "I can't find this user's name in my environment". That is apparently true, depending on which method is used to query the environment.. When I run id in one of those stalled sessions, I get:

uid=501 ...

In a fresh session, it is

uid=501(selkovjr)

Same result from whoami.

However, the $USER variable still has my username even in stalled sessions.

The issue is resolved by simply restarting the session through the Shell menu. I like how that works because my screen buffers remain intact and I can see what I was doing in each session before it got disrupted.

Why is this issue closed when it is not fixed? I still experiencing this issue with iterm build 2.9.20160510.

@selkovjr Does it happen for you with newer versions?

@depressiveRobot What are the steps to reproduce for you?

Status changed to reopened

@gnachman Same as for the other reporters: I closed the lid to put it to sleep and later to error occurred. I must admit, that this error only happened once until now.

@depressiveRobot Can you reproduce it consistently or does it happen only rarely when you close the lid and it sleeps?

Sorry for the delayed response; I was going to give it more testing but my mac took a break and is now waiting for a new logic board.

For me, it only happened once, although with lasting after-effects. I had many sessions open before I closed the lid, so when it woke up in that broken state, I could diff the environments between the disconnected sessions and the new ones. The only observed difference was was that my uid did not resolve to a user name in the disconnected ones.

Then, I think I continued to run the same "faulty" version for about three weeks putting it to sleep every day, without problems. Then I rebooted and installed the latest iterm; saw no problems for several more days and then my GPU melted down. So even though my experiment has been inconclusive, I can definitely say the problem was intermittent. No idea what could have caused it. I am suspecting, it must be something in with the way ssh takes down connections. What else? We start with a good environment, run ssh, go to sleep and wake up (that should not matter), the ssh sees a stale socket and bails out, presenting us with a modified environment in the parent shell. Is it possible for a shell's environment (bash in my case) to be modified from without? If not, ssh is the likely culprit.

Thanks @selkovjr. No, and environment can't be changed by a parent job. If someone can repro consistently I'll try to drill down on it, otherwise it's tough to do more than I've already done.

I'm getting this error too. There was some crash that logged out my user on OS X, then I couldn't use it after logging back in.

The same thing happened to me at least twice when something crashed my whole OS X user session and threw me back to the login screen. iTerm2 (Build 3.0.10) recreates all the windows and they seem to work apart from the No user exists for uid 501 error I get when I run SSH. Also the id -un prints only the UID (without the name): 501

To reproduce this you should be able to open an iTerm2 window with at least one session, kill WindowServer and log back in. The iTerm2 session should still be active and you should be able to reproduce the behaviour.

I have the same error happening when OSX runs out of memory and throws to login screen. When iTerm2 is restored git commands don't work.

The full message:

$git pull
No user exists for uid 502
fatal: Could not read from remote repository.

Please make sure you have the correct access rights
and the repository exists.

Same here, after my Sierra session crashed the GitX binary didn't work and I got “No user exists for uid 50x” messages when running git commands.

I tried sudo killall - WindowServer and then logging back in. Things are a little funky, like whois says 501 instead of my user name. Problem is, I don't think I can detect when this happens. Even if I could, I'm not sure that killing running sessions would be that great. You could still save your data even if you aren't able to fork new processes in the half-broken shell…

mentioned in issue #5414 (closed)

I've noticed it recently. The pattern I've noticed is when I 'undock' my laptop from a monitor. It could also happen to be sleeping at the time.

Opening a new terminal window resolves it for me, but existing terminal windows have the problem.

Happened to me again today. Using the Shell Menu Restart Session in each broken tab works, but hoping for a better solution.

Happened to me.

$git pull
No user exists for uid 502
fatal: Could not read from remote repository.

Please make sure you have the correct access rights
and the repository exists.

I open a new tab or Restart Session to resolve it.

But I need a better solution.

What about a "Restart All Sessions" option? Seems like the only thing to do in this case.

Happening to me as well. Not sure what I did. Guessing it has something to do with my system rebooting itself after a Panic.

I faced this issue today. Had to restart my session to get it working.

any word on @gnachman on this issue?

Several of us in my office suffer from the same issue

Someone who can reproduce this, please try this build and let me know if it makes any difference: http://iterm2.com/adhocbuilds/iTerm2-3_1_20170424_220422-adhoc.zip

I just got this in the nightly build (haven't seen it in a while). I force quit and started the adhoc build so it would use the same sessions, and it "restored" the sessions with the same "No user exists for uid 501" error with git.

@asmeurer That build kills the sessions when you log off. So the question is whether it would have prevented it. Since people don't want to remain on old builds, I've added an advanced setting to tomorrow's nightly build in commit 316b855. Turn on Prefs > Advanced > Kill sessions on logout and see if the problem goes away or other bad behavior appears after logging out.

@gnachman I've been seeing this issue every once in a while when my system does what I would call an unexpected "soft reboot". Once every few weeks, in the middle of using my machine, the screen fades to black and then to the login screen (power is not lost, and the machine does not actually reboot). Once I login, all of my previously-running applications start fresh (Chome gives the "Chrome was not shut down properly" message, etc). I'm not sure what triggers this "soft reboot", but it may be related to never shutting down or rebooting manually. I haven't been able to reproduce it by attempting to overload my (16GB of) RAM or (i7 quad-core) CPU.

All of that background to say that a few weeks ago, I installed the build you posted above, and I just now experienced a "soft reboot" for the first time since then. iTerm appears to have "restored" all sessions as it had been doing previously. I tried a "git fetch" in one of the terminals, and I received the same No user exists for uid 501 error message. Thus, the build you posted does not seem to make a difference.

To confirm, I double-checked iTerm2 > About iTerm2, and the listed version is Build 3.1.20170424.

mentioned in issue #5774

@airdrummingfool Can you try turning on Prefs > Advanced > Kill sessions on logout and see if that helps with the soft reboots?

@gnachman I didn't have that option in the build I was on (3.1.20170424), so I've updated to 3.1.beta.4 and enabled it. I'll post back once I experience another soft reboot. Thanks!

@gnachman I just experienced a soft reboot, and the Kill sessions on logout setting did not kick in. All of my sessions were restored, including a running Python script, an open SSH session, and a long-running gulp command. I double-checked, and the setting was set to Yes, so whatever is happening to me must not constitute a "logout".

I thought I had toggled debug logging, but apparently not, as when I tried to toggle it again it said "Please reproduce the bug and then toggle debug logging again to save the log.". The debug logging thing is a bit confusing to me; there doesn't seem to be a way to know whether or not it is on and currently collecting log data. I just found out about the advanced setting to enable logging on iTerm start, so hopefully I'll be able to get you a log the next time this happens :)

EDIT: Just experienced another soft reboot. Unfortunately the debug log seems to get reset each time iTerm starts up, so the log begins after the system has recovered and restarted the applications. If you think it could still be useful, let me know of a good way to get it to you - it's ~4MB zipped and has potentially sensitive information in it (so I don't want to post it here/publicly).

Based on what you saw, it looks like the system killed iTerm2 on logout. Tomorrow's nightly build will write out the debug log on app termination (commit d0b51a2), although I'd be surprised if it's informative. Console.app logs might be more useful, although they tend to be full of useless spam.

Kind of sounds like a system daemon is crashing, as described here: https://www.cnet.com/news/managing-unexpected-log-outs-in-os-x/

Session restoration doesn't kill terminals on logout

Related issues