[[inputs.ntpq]] works, but produce errors in the log "E! Error ntpq: parsing int: u" #2386

sunserg · 2017-02-09T16:54:53Z

Bug report

telegraf.service log shows the error every 10s: "E! Error ntpq: parsing int: u"

osboxes@osboxes:~$ sudo systemctl status telegraf.service
● telegraf.service - The plugin-driven server agent for reporting metrics into InfluxDB
   Loaded: loaded (/lib/systemd/system/telegraf.service; enabled; vendor preset: enabled)
   Active: active (running) since Thu 2017-02-09 13:05:04 GMT; 3h 40min ago
     Docs: https://github.com/influxdata/telegraf
 Main PID: 2237 (telegraf)
   CGroup: /system.slice/telegraf.service
           └─2237 /usr/bin/telegraf -config /etc/telegraf/telegraf.conf -config-directory /etc/telegraf/telegraf.d

Feb 09 16:44:30 osboxes telegraf[2237]: 2017-02-09T16:44:30Z E! Error ntpq: parsing int: u
Feb 09 16:44:40 osboxes telegraf[2237]: 2017-02-09T16:44:40Z E! Error ntpq: parsing int: u
Feb 09 16:44:50 osboxes telegraf[2237]: 2017-02-09T16:44:50Z E! Error ntpq: parsing int: u
Feb 09 16:45:00 osboxes telegraf[2237]: 2017-02-09T16:45:00Z E! Error ntpq: parsing int: u
Feb 09 16:45:10 osboxes telegraf[2237]: 2017-02-09T16:45:10Z E! Error ntpq: parsing int: u
Feb 09 16:45:20 osboxes telegraf[2237]: 2017-02-09T16:45:20Z E! Error ntpq: parsing int: u
Feb 09 16:45:30 osboxes telegraf[2237]: 2017-02-09T16:45:30Z E! Error ntpq: parsing int: u
Feb 09 16:45:40 osboxes telegraf[2237]: 2017-02-09T16:45:40Z E! Error ntpq: parsing int: u
Feb 09 16:45:50 osboxes telegraf[2237]: 2017-02-09T16:45:50Z E! Error ntpq: parsing int: u
Feb 09 16:46:00 osboxes telegraf[2237]: 2017-02-09T16:46:00Z E! Error ntpq: parsing int: u
osboxes@osboxes:~$

Relevant telegraf.conf:

# # Get standard NTP query metrics, requires ntpq executable.
[[inputs.ntpq]]
#   ## If false, set the -n ntpq flag. Can reduce metric gather time.
  dns_lookup = true

System info:

Linux osboxes 4.4.0-59-generic #80-Ubuntu SMP Fri Jan 6 17:47:47 UTC 2017 x86_64 x86_64 x86_64 GNU/Linux
telegraf v. 1.2.1-1

Steps to reproduce:

enable ntpq input in telegraf.conf
restart telegraf.service

Expected behavior:

no errors

Actual behavior:

errors

Additional info:

output of ntpq is:

osboxes@osboxes:~$ ntpq -p 
     remote           refid      st t when poll reach   delay   offset  jitter
==============================================================================
#192.168.2.101   194.109.6.2      3 u  236 1024  377    0.598    4.504  11.352
+194.109.22.18   193.67.79.202    2 u  317 1024  377    5.177    1.202   0.202
 0.ubuntu.pool.n .POOL.          16 p    -   64    0    0.000    0.000   0.000
 1.ubuntu.pool.n .POOL.          16 p    -   64    0    0.000    0.000   0.000

The text was updated successfully, but these errors were encountered:

bascht · 2017-07-03T14:51:54Z

Same here. More output, if it helps debugging:

$ ntpq -p 
     remote           refid      st t when poll reach   delay   offset  jitter
==============================================================================
 0.ubuntu.pool.n .POOL.          16 p    -   64    0    0.000    0.000   0.000
 1.ubuntu.pool.n .POOL.          16 p    -   64    0    0.000    0.000   0.000
 2.ubuntu.pool.n .POOL.          16 p    -   64    0    0.000    0.000   0.000
 3.ubuntu.pool.n .POOL.          16 p    -   64    0    0.000    0.000   0.000
 ntp.ubuntu.com  .POOL.          16 p    -   64    0    0.000    0.000   0.000
*char-ntp-pool.c .SHM.            1 u  579 1024  377   21.200    0.296   0.362
+37.58.57.238 (d 192.53.103.103   2 u   10 1024  377    1.748    0.373   0.101
+srv24.globale-g 131.188.3.222    2 u  682 1024  377    5.637    0.611   0.295
-re.uni-paderbor .DCF.            1 u  268 1024  377   21.700    4.553   1.030
-tms-proxy.smart 193.67.79.202    2 u  688 1024  377    5.101   -0.233   0.303

dsalbert · 2017-07-17T16:02:15Z

I think this error is related to vary number of columns provided by ntpq command for some cases like e.g.

+37.58.57.238 (d 192.53.103.103   2 u   10 1024  377    1.748    0.373   0.101

Character that is splitting columns is a regular space, so adjusting fields := strings.Fields(line) function to use only TAB is not possible.

My propose is a little bit "hacky":

		fields := strings.Fields(line)

		if len(fields) < 2 {
			continue
		}

		// In case when, there will be more columns that desired (10)		
		desiredColumns := 10
		if len(fields) > desiredColumns {
			var (
				deleteStop = (len(fields) - desiredColumns) + 1
				firstField = strings.Join(fields[0:deleteStop], "")
			)

			fields = append(fields[:1], fields[deleteStop:]...)
			fields[0] = firstField
		}

What do you think?

danielnelson · 2017-07-17T19:32:59Z

I wonder if ought to first enable -w before spending time fixing the current format. It will allow us to report the full dns entry instead of a potentially truncated value. This will change the output to be like:

     remote           refid      st t when poll reach   delay   offset  jitter
==============================================================================
 0.debian.pool.ntp.org
                 .POOL.          16 p    -   64    0    0.000    0.000   0.000
 1.debian.pool.ntp.org
                 .POOL.          16 p    -   64    0    0.000    0.000   0.000
 2.debian.pool.ntp.org
                 .POOL.          16 p    -   64    0    0.000    0.000   0.000
 3.debian.pool.ntp.org
                 .POOL.          16 p    -   64    0    0.000    0.000   0.000
#kapu.ruselabs.com
                 200.98.196.212   2 u   56   64    1   87.060    0.095   0.388
#triton.ellipse.net
                 216.229.0.179    2 u   55   64    1   69.776  -17.777   0.492
#ns1.oninit.net  198.55.111.5     3 u   55   64    1   53.744    1.926   0.592
+x.ns.gin.ntt.net
                 249.224.99.213   2 u   29   64    3   13.428   -0.100   0.674
-138.68.46.177 (au.kashra.pictures)
                 90.187.7.5       3 u   29   64    3   14.746   -1.147   1.262
-services.quadranet.com
                 45.79.111.114    3 u   29   64    3   24.820   -3.205   0.653
-ntp-3.jonlight.com
                 66.220.9.122     2 u   28   64    3   14.785    2.040   1.326
-chimera.bufferoverflow.xyz
                 128.252.19.1     2 u   28   64    3   63.300   -5.883   0.568
#eterna.binary.net
                 216.229.0.50     3 u   29   64    3   70.741  -12.018   0.948
-time-a.timefreq.bldrdoc.gov
                 .NIST.           1 u   30   64    3   84.074   -2.977   0.548
+ntp.idealab.com 208.76.2.12      2 u   28   64    3   24.207    0.570   0.713
*utcnist2.colorado.edu
                 .NIST.           1 u   27   64    3   42.894   -1.280   0.436
-ec2-75-101-249-126.compute-1.amazonaws.com
                 138.39.23.13     2 u   25   64    3   80.911   -0.274   0.740
#96.226.123.195 (smtp.visionnet.us)
                 216.229.0.179    2 u   27   64    3   53.842  -12.868   0.581
#lithium.constant.com
                 18.26.4.105      2 u   27   64    3   83.253    2.217   0.638
-209.208.79.69   209.51.161.238   2 u   27   64    3   84.091   -1.290   0.419

From man ntpq:

              Display the full value of the 'remote' value.  If this requires
              more than 15 characters, display the full value, emit a newline,
              and continue the data display properly indented on the next
              line.

If we do this the parser will need to change quite a bit, but I think I would either try tokenizing with a regex or possibly iterating right to left.

dsalbert · 2017-07-17T20:42:35Z

I do not have this option in my binary. Which version of ntpq are you using?
If we know that we need to remove from a string an expression that starts from ( and end up with a space, what about just using \((\s|\S+) regexp, on whole stream or on each line?
e.g.

	reg, _ := regexp.Compile("\\((\\s|\\S+)")
	output1 := reg.ReplaceAllString("#96.226.123.195 (smtp.visionnet.us) 16.229.0.179    2 u   27   64    3   53.842 -12.868 0.581", "")
	output2 := reg.ReplaceAllString("#96.226.123.195 (s 216.229.0.179    2 u   27   64    3   53.842  -12.868   0.581", "")
	output3 := reg.ReplaceAllString("#96.226.123.195 ( 216.229.0.179    2 u   27   64    3   53.842  -12.868   0.581", "")

result:

#96.226.123.195  16.229.0.179    2 u   27   64    3   53.842 -12.868 0.581
#96.226.123.195  216.229.0.179    2 u   27   64    3   53.842  -12.868   0.581
#96.226.123.195 216.229.0.179    2 u   27   64    3   53.842  -12.868   0.581

danielnelson · 2017-07-17T21:48:02Z

I guess -w is too new, both 4.2.8p10 in debian stretch and gentoo have it, but 4.2.6p5 in Centos 7 does not.

Regex replacement on the whole line sounds good, if you want to line up the whitespace for no reason other than because you are crazy like me you could do:

`\([^\s]*`

dsalbert · 2017-07-18T10:39:35Z

As a proof that I am, I've used \([\S]*;>

- added test cases for SHM remotes; - slightly changed the fix for influxdata#2386 to improve and fix the parsing of SHM remotes.

sparrc added bug unexpected problem or unintended behavior help wanted Request for community participation, code, contribution labels Feb 9, 2017

sparrc added this to the Future Milestone milestone Feb 9, 2017

danielnelson removed this from the Future Milestone milestone Jun 14, 2017

dsalbert pushed a commit to dsalbert/telegraf that referenced this issue Jul 18, 2017

add regular expression to tackle a problem influxdata#2386

e8ca023

dsalbert mentioned this issue Jul 18, 2017

Fix ntpq parse issue when using dns_lookup #3026

Merged

3 tasks

danielnelson added this to the 1.4.0 milestone Jul 18, 2017

danielnelson closed this as completed Jul 18, 2017

danielnelson modified the milestones: 1.3.5, 1.4.0 Jul 25, 2017

jawi mentioned this issue Aug 24, 2017

NTP plugin incorrectly parses shared memory remotes #3162

Closed

jawi pushed a commit to jawi/telegraf that referenced this issue Aug 24, 2017

Fixed influxdata#3162 fix parsing of SHM remotes

54ffb4f

- added test cases for SHM remotes; - slightly changed the fix for influxdata#2386 to improve and fix the parsing of SHM remotes.

jawi mentioned this issue Aug 24, 2017

Fixed #3162 fix parsing of SHM remotes #3163

Merged

3 tasks

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

[[inputs.ntpq]] works, but produce errors in the log "E! Error ntpq: parsing int: u" #2386

[[inputs.ntpq]] works, but produce errors in the log "E! Error ntpq: parsing int: u" #2386

sunserg commented Feb 9, 2017

bascht commented Jul 3, 2017

dsalbert commented Jul 17, 2017

danielnelson commented Jul 17, 2017

dsalbert commented Jul 17, 2017 •

edited

Loading

danielnelson commented Jul 17, 2017

dsalbert commented Jul 18, 2017

[[inputs.ntpq]] works, but produce errors in the log "E! Error ntpq: parsing int: u" #2386

[[inputs.ntpq]] works, but produce errors in the log "E! Error ntpq: parsing int: u" #2386

Comments

sunserg commented Feb 9, 2017

Bug report

Relevant telegraf.conf:

System info:

Steps to reproduce:

Expected behavior:

Actual behavior:

Additional info:

bascht commented Jul 3, 2017

dsalbert commented Jul 17, 2017

danielnelson commented Jul 17, 2017

dsalbert commented Jul 17, 2017 • edited Loading

danielnelson commented Jul 17, 2017

dsalbert commented Jul 18, 2017

dsalbert commented Jul 17, 2017 •

edited

Loading