User login
QR Code
Syndicate
ntpd and bad file descriptor
Wed, 2007-02-14 13:17 — johnp
Some days ago I noticed that ntpd wrote an error to the syslog. The message read:
Feb 13 12:11:06 servername ntpd[4415]: sendto(192.168.100.100) (fd=-1): Bad file descriptor
At first I believed there was an error in the configuration file, so I reinstalled the package to be sure that I hadn't misconfigured anything. Looking at the running processes I noticed that there were two ntp daemons running. So I stopped the ntpd using
/etc/init.d/ntp stop
Listing active processes once again revealed that process 4415, in this case the ntpd daemon was still running, but now alone. By killing the process and then start ntp in the usual fashion the problem with bad file descriptor was solved.
Comments
Excellent bro. This helped
Excellent bro. This helped me!
Hi, It helped me solve
Hi,
It helped me solve similar issue in rhel5 U2. Can you tell me the step by step process to do public key encryption in ntp server as I have done yet only symmetric encryption only.
mail : jeevanullas at gmail dot com
Thanks in advance
Deependra
Public key encryption
Sorry, but I have to inform you that I haven't done any asymmetric encryption for the use with NTP. have you checked http://www.ntp.org or http://www.eecis.udel.edu/~mills/ntp/html/authopt.html#pub?
// John
Thanks!
Haha... I thought I was going mad, but indeed: there were two ntpd processes. Thanks dude!
Thank you, johnp
This was very helpful
Nice reminder, thanks!
If I just would have thought what I was doing :p
Thank you!
Found this via Google and it clearly matched the issue when I checked. Very grateful you took the time to post the information.
Solved also on CentOS 5.4
Thank you for you effort, it helped me also :)
Solved also on CentOS 5.4
Thank you for your very helpfull post, helped me to solve the problem even on a CentOS 5.4 machine :)
Solved also on CentOS 5.4
Thank you helped me also on CentOS 5.4
But was that an ntp server...
I have the same issue (running RHEL4u6), and after reboot two ntp daemons will show up. Both reference the same pid file. I'll figure out which one is *not* affected by service ntpd stop, and kill that one. I'll restart the service and all is fine. Note that this box is the time server on an isolated network.
Now, the instant a client attempts to poll for time - bang! Two instances again. I'll work through it, and if I fugure out what it is before someone else does, I'll post it.
Interesting
Please let us know if you find any clues what may cause this.
// John
Thanks! Same problem here
Thanks! Same problem here (CentOS 5.6) and your advice worked flawlessly. This is what I like about Linux—the community aspect. People don't sit around discussing what caused this week's Windows BSOD on the net. At least, I hope not...
Post new comment