| RE: [chrony-users] chrony-3.4 bug? |
[ Thread Index |
Date Index
| More chrony.tuxfamily.org/chrony-users Archives
]
Hi Miroslav.
Correct - it doesn’t start as a system service. The pidfile is specified within my chrony.conf file so I wouldn’t expect that it will be responsible. I also didn’t upgrade from a previous version of chrony - this is a brand new deployment (still under test). I downloaded the 3.4 source code bundle and compiled it on a RHEL 7 VM.
When I try to start it with the -n switch in place I get the following messages:
>sudo systemctl restart chronyd.service
(after a long timeout)
Job for chronyd.service failed because a timeout was exceeded. See "systemctl status chronyd.service" and "journalctl -xe" for details.
>sudo systemctl status chronyd.service
● chronyd.service - Chrony
Loaded: loaded (/etc/systemd/system/chronyd.service; enabled; vendor preset: enabled)
Active: activating (start) since Fri 2018-11-02 13:26:47 EDT; 32s ago
Main PID: 20946 (code=exited, status=0/SUCCESS); : 20999 (chronyd)
CGroup: /system.slice/chronyd.service
└─20999 /var/chrony/sbin/chronyd -4 -n -u chrony -f /var/chrony/etc/chrony.conf -l /var/chrony/log/chrony.log
Nov 02 13:26:47 <hostname> systemd[1]: chronyd.service holdoff time over, scheduling restart.
Nov 02 13:26:47 <hostname> systemd[1]: Starting Chrony...
>sudo journalctl -xe
-- Unit chronyd.service has begun shutting down.
Nov 02 13:25:16 <hostname> systemd[1]: Starting Chrony...
-- Subject: Unit chronyd.service has begun start-up
-- Defined-By: systemd
-- Support: http://lists.freedesktop.org/mailman/listinfo/systemd-devel
--
-- Unit chronyd.service has begun starting up.
Nov 02 13:26:46 <hostname> systemd[1]: chronyd.service start operation timed out. Terminating.
Nov 02 13:26:46 <hostname> systemd[1]: Failed to start Chrony.
-- Subject: Unit chronyd.service has failed
-- Defined-By: systemd
-- Support: http://lists.freedesktop.org/mailman/listinfo/systemd-devel
--
-- Unit chronyd.service has failed.
--
-- The result is failed.
Nov 02 13:26:46 <hostname> systemd[1]: Unit chronyd.service entered failed state.
Nov 02 13:26:46 <hostname> systemd[1]: chronyd.service failed.
Nov 02 13:26:46 <hostname> polkitd[621]: Unregistered Authentication Agent for unix-process:20981:44148827 (system bus name :1.10355, object path /org/freedesktop/Polic
Nov 02 13:26:47 <hostname> systemd[1]: chronyd.service holdoff time over, scheduling restart.
Nov 02 13:26:47 <hostname> systemd[1]: Starting Chrony...
-- Subject: Unit chronyd.service has begun start-up
-- Defined-By: systemd
-- Support: http://lists.freedesktop.org/mailman/listinfo/systemd-devel
--
-- Unit chronyd.service has begun starting up.
Note that chrony never actually starts - it just times out.
Here is the unit file (assuming you mean the chronyd.service file that I created):
_______________________
[Unit]
Description=Chrony
After=network.target
[Service]
ExecStart=/var/chrony/sbin/chronyd -4 -n -u chrony -f /var/chrony/etc/chrony.conf -l /var/chrony/log/chrony.log
Type=forking
PIDFile=/var/chrony/var/run/chrony.pid
Restart=always
[Install]
WantedBy=default.target
_______________________
....and here is the chrony.conf file that I created (IPs removed):
_______________________
#### TIME SOURCE SERVERS
## Define servers to draw time from
## Reduce polling interval to increase accuracy
## minpoll 2 equates to 4s
## maxpoll 3 equates to 8s
## iburst sends the first 4 requests at an interval of 2s or less
## Define the time source
## Use ONLY the time source that exists within the same site
server <IP1> minpoll 2 maxpoll 3 iburst version 4 maxdelay 1 trust prefer require
## Note that it is recommended for security purposes that peering arrangements not be used
## Instead, it is recommended that each peer specify a client/server arrangement with one another
## Use ONLY the same site <redacted> appliance
server <IP2> minpoll 2 maxpoll 3 iburst key 312 version 4 maxdelay 1 trust prefer require
## Use ONLY the same site other chronyd servers
server <IP3> minpoll 2 maxpoll 3 iburst key 311 version 4 maxdelay 1 trust prefer require
#### TIME ADJUSTMENT CONSTRAINTS
## Limit the number of sources used (default)
combinelimit 3
## Do not update the local clock without having at least 2 sources considered as selectable
minsources 2
## Select slew for leapsecond treatment
leapsecmode slew
## Set timezone for leap seconds
leapsectz right/UTC
## Step the local system clock if it is off by more than 0.01s in the first 3 clock updates
makestep 0.01 3
## Set the max allowed clock offset correction on a clock update
maxchange 1000 1 -1
## Set max update skew to 5 ppm (good LAN)
maxupdateskew 5
#### LOCAL SERVER SPECIFICS
## Define the address to which chrony will bind as a client
## Not applicable to server configuration
#bindacqaddress <IP_address>
## Define the address to which chrony will bind as a server
bindaddress <local_IP>
## Define location of measurement history file
dumpdir /var/chrony/data/measurement_hist
## Store drift rate information in this file
driftfile /var/chrony/data/drift
## Copy system time to Real-Time Clock (RTC)
rtcsync
## NOTE That hardware timestamping is not supported on the network interface cards so will not be used
## Keys
keyfile /var/chrony/keys/chrony.keys
## Lock chrony in memory such that it will not swap out to disk (results in lower and more consistent latency)
lock_all
## PID file
pidfile /var/chrony/var/run/chrony.pid
#### SECURITY
## Deny all remote command attempts
# cmddeny all
## 181028 - allow command-level access from other NTP servers (originally required for testing purposes)
cmdallow <other_IP_1>
cmdallow <other_IP_2>
cmdallow <other_IP_3>
cmdallow <other_IP_4>
cmdallow <other_IP_5>
cmdallow <other_IP_6>
## 181031 - Define the address to which chrony will bind for commands
bindcmdaddress <local_IP>
## 181031 - Open the command port (default port #)
cmdport 323
## Specify user to run chrony as
user chrony
#### CONTROLLING CLIENT ACCESS
## Allow queries from anyone (restricted on the firewall)
allow all
## Set max memory allocated to tracking client access
clientloglimit 1048576
## Specify port number
port 123
## Rate limit client NTP queries
ratelimit interval 2
#### LOGGING
## Define log directory
logdir /var/chrony/log
## Define what to log
log measurements statistics tracking
## Define threshold for system clock adjustments to be logged (default)
logchange 1
_______________________
Interestingly enough, the chronyc output is what I needed and it appears to work fine with the -n switch. So if this is a bug it is not holding me up at the moment.
Thanks!
Dan LeBlanc
-----Original Message-----
From: Miroslav Lichvar [mailto:mlichvar@xxxxxxxxxx]
Sent: November-01-18 5:44 AM
To: chrony-users@xxxxxxxxxxxxxxxxxxxx
Subject: Re: [chrony-users] chrony-3.4 bug?
On Wed, Oct 31, 2018 at 11:55:29PM +0000, LeBlanc, Daniel James wrote:
> Hi All.
>
> I have discovered that chrony-3.4 will not startup on RHEL 7 if I pass it the -n option. It times out. My executable start string is as follows:
>
> /var/chrony/sbin/chronyd -4 -u chrony -f /var/chrony/etc/chrony.conf -l /var/chrony/log/chrony.log
>
> If I add the -n after the -4 or replace the -4 with -n, chronyd fails to start. Has anyone else successfully used the -n flag?
It doesn't start as a systemd service? Can you please post the chrony
unit file and chrony config?
There was one change in chrony-3.4 that could be related. The default
pidfile has moved from /var/run to /var/run/chrony. So, you may need
to update the unit file or chrony config.
--
Miroslav Lichvar
--
To unsubscribe email chrony-users-request@xxxxxxxxxxxxxxxxxxxx
with "unsubscribe" in the subject.
For help email chrony-users-request@xxxxxxxxxxxxxxxxxxxx
with "help" in the subject.
Trouble? Email listmaster@xxxxxxxxxxxxxxxxxxxx.
--
To unsubscribe email chrony-users-request@xxxxxxxxxxxxxxxxxxxx
with "unsubscribe" in the subject.
For help email chrony-users-request@xxxxxxxxxxxxxxxxxxxx
with "help" in the subject.
Trouble? Email listmaster@xxxxxxxxxxxxxxxxxxxx.