[ZBX-15102] Zabbix 4.0.1 proxy housekeeper process constantly 100% busy. Created: 2018 Oct 31  Updated: 2018 Oct 31  Resolved: 2018 Oct 31

Status: Closed
Project: ZABBIX BUGS AND ISSUES
Component/s: Proxy (P)
Affects Version/s: 4.0.1
Fix Version/s: None

Type: Problem report Priority: Minor
Reporter: Rai Assignee: Unassigned
Resolution: Commercial support required Votes: 0
Labels: 4.0.1, proxy,
Remaining Estimate: Not Specified
Time Spent: Not Specified
Original Estimate: Not Specified
Environment:

Centos 7


Attachments: PNG File 100.png    

 Description   

With no apparent reason Zabbix 4.0.1 proxy housekeeper process is constantly 100% busy.

Screen attached to this ticket.

zabbix_proxy logs looks good:

  3802:20181030:154732.485 proxy #5 started [housekeeper #1]
  3802:20181030:161732.739 executing housekeeper
  3802:20181030:161759.804 housekeeper [deleted 9547964 records in 27.059241 sec, idle for 1 hour(s)]
  3802:20181030:171800.241 executing housekeeper
  3802:20181030:171816.615 housekeeper [deleted 6976314 records in 16.366344 sec, idle for 1 hour(s)]
  3802:20181030:181817.066 executing housekeeper
  3802:20181030:181836.779 housekeeper [deleted 7029910 records in 19.703083 sec, idle for 1 hour(s)]
  3802:20181030:191837.209 executing housekeeper
  3802:20181030:191856.242 housekeeper [deleted 7083754 records in 19.024195 sec, idle for 1 hour(s)]
  3802:20181030:201856.691 executing housekeeper
  3802:20181030:201914.101 housekeeper [deleted 7096330 records in 17.397830 sec, idle for 1 hour(s)]
  3802:20181030:211914.539 executing housekeeper
  3802:20181030:211931.967 housekeeper [deleted 7153524 records in 17.421797 sec, idle for 1 hour(s)]
  3802:20181030:221932.390 executing housekeeper
  3802:20181030:221950.101 housekeeper [deleted 7078202 records in 17.702982 sec, idle for 1 hour(s)]
  3802:20181030:231950.542 executing housekeeper
  3802:20181030:232009.387 housekeeper [deleted 7041819 records in 18.837928 sec, idle for 1 hour(s)]
  3802:20181031:002009.834 executing housekeeper
  3802:20181031:002030.239 housekeeper [deleted 7152769 records in 20.395927 sec, idle for 1 hour(s)]
  3802:20181031:012030.697 executing housekeeper
  3802:20181031:012046.850 housekeeper [deleted 7146650 records in 16.145200 sec, idle for 1 hour(s)]
  3802:20181031:022047.301 executing housekeeper
  3802:20181031:022106.667 housekeeper [deleted 7154168 records in 19.352890 sec, idle for 1 hour(s)]
  3802:20181031:032107.105 executing housekeeper
  3802:20181031:032126.438 housekeeper [deleted 7147777 records in 19.325092 sec, idle for 1 hour(s)]
  3802:20181031:042126.886 executing housekeeper
  3802:20181031:042145.189 housekeeper [deleted 7170236 records in 18.294948 sec, idle for 1 hour(s)]
  3802:20181031:052145.618 executing housekeeper
  3802:20181031:052206.535 housekeeper [deleted 7174243 records in 20.909850 sec, idle for 1 hour(s)]
  3802:20181031:062206.968 executing housekeeper
  3802:20181031:062224.236 housekeeper [deleted 7167066 records in 17.258151 sec, idle for 1 hour(s)]
  3802:20181031:072224.680 executing housekeeper
  3802:20181031:072243.566 housekeeper [deleted 7165431 records in 18.879511 sec, idle for 1 hour(s)]
  3802:20181031:082244.017 executing housekeeper
  3802:20181031:082304.978 housekeeper [deleted 7125541 records in 20.951898 sec, idle for 1 hour(s)]
  3802:20181031:092305.416 executing housekeeper
  3802:20181031:092322.863 housekeeper [deleted 6989932 records in 17.440464 sec, idle for 1 hour(s)]
  3802:20181031:102323.303 executing housekeeper
  3802:20181031:102342.878 housekeeper [deleted 6947359 records in 19.568866 sec, idle for 1 hour(s)]
  3802:20181031:112343.323 executing housekeeper
  3802:20181031:112401.003 housekeeper [deleted 6947317 records in 17.670088 sec, idle for 1 hour(s)]


 Comments   
Comment by Rai [ 2018 Oct 31 ]

One spike to 0% you see at the graph happened while we upgraded server and proxy from 4.0.0 to 4.0.1 version.

Comment by Ingus Vilnis [ 2018 Oct 31 ]

In Configuration -> Hosts for the proxy is the host monitored by the proxy itself or by server directly?

Comment by Vladislavs Sokurenko [ 2018 Oct 31 ]

could you please show output of:
watch -tn 0.2 'ps -fC zabbix_proxy | grep housekeeper'
watch -tn 0.2 'ps -fC zabbix_server | grep housekeeper'

As ingus.vilnis mentioned there might be another process monitored.

Comment by Rai [ 2018 Oct 31 ]

Proxy is monitored by server directly.

On zabbix proxy host:

 zabbix    3802  3727  0 Oct30 ?        00:00:03 /usr/sbin/zabbix_proxy: housekeeper [deleted 6976106 records in 19.212090 sec, idle for 1 hour(s)]

On zabbix server:

zabbix    4824  0.1  0.6 3431480 398120 ?      S    Oct30   1:36  \_ /usr/sbin/zabbix_server: housekeeper [removing deleted items data]

I disabled the housekeeper on zabbix server and will look if this helps to solve the problem on zabbix proxy.

Comment by Ingus Vilnis [ 2018 Oct 31 ]

Proxy must be monitored by itself, not by server. What you are seeing is actually a problem with housekeeper on Zabbix server. 

Disabling the housekeeper on server will stop deleting old data and eventually you will run out of disk space and your monitoring will crash. Consider to implement partitioning of history* and trends* tables as soon as possible. 

Not a bug in Zabbix. 

Comment by Rai [ 2018 Oct 31 ]

I already have partitions on server. This was a surprise that housekeeper was running on a server at the moment.

Will change proxy host settings and see if this helps.

Comment by Vladislavs Sokurenko [ 2018 Oct 31 ]

No indication of a bug, closing as support required.

Generated at Tue Apr 23 18:23:28 EEST 2024 using Jira 9.12.4#9120004-sha1:625303b708afdb767e17cb2838290c41888e9ff0.