[ZBX-8120] Duplicated history values with zeroes after 1 second Created: 2014 Apr 18  Updated: 2017 Oct 24  Resolved: 2017 Oct 24

Status: Closed
Project: ZABBIX BUGS AND ISSUES
Component/s: Proxy (P), Server (S)
Affects Version/s: 2.2.2, 2.2.3
Fix Version/s: None

Type: Incident report Priority: Critical
Reporter: DKoshelev Assignee: Unassigned
Resolution: Duplicate Votes: 1
Labels: duplicates, hiload
Remaining Estimate: Not Specified
Time Spent: Not Specified
Original Estimate: Not Specified

Attachments: PNG File chart.php.png    
Issue Links:
Duplicate
duplicates ZBX-12281 SNMPv2 poller and/or proxy creates du... Closed

 Description   

We have found strange behaior in zabbix 2.2.2 and 2.2.3. Some graphs have pikes to zero.
In mysql it looks like:

mysql> select * from history where itemid="63512";
+--------+------------+----------+-----------+
| itemid | clock      | value    | ns        |
+--------+------------+----------+-----------+
|  63512 | 1397802752 | 247.0409 | 526085751 | 
|  63512 | 1397803112 | 262.5530 | 450004908 | 
|  63512 | 1397803113 |   0.0000 | 450004908 | 
|  63512 | 1397803232 | 264.6991 |  82818278 | 
|  63512 | 1397803352 | 262.0739 | 693831403 | 
|  63512 | 1397803472 | 257.6663 | 255432410 | 
|  63512 | 1397803473 |   0.0000 | 255432410 | 
|  63512 | 1397803592 | 259.0060 | 905185825 | 
|  63512 | 1397803712 | 266.8781 | 547834755 | 
|  63512 | 1397803832 | 258.2837 | 121761075 | 
|  63512 | 1397803833 |   0.0000 | 121761075 | 
+--------+------------+----------+-----------+

Zabbix server get correct data from proxy:

2014-04-18 10:50:45.000819 [30538] recv_data_from_proxy() [{mysql> select * from history where itemid="63512";
.....
 {
                        "host":"1080232",
                        "key":"blah-blah",
                        "clock":1397803832,
                        "ns":121761075,
                        "value":"243375193.000000"},
......

But after it looks like:

2014-04-18 10:50:46.000046 [30528] query [txnlev:1] [insert into history (itemid,clock,ns,value) values .... (63512,1397803832,121761075,258.283733) ....
....
2014-04-18 10:50:46.000081 [30528] In zbx_vc_add_value() itemid:63512 value_type:0 timestamp:1397803832.121761075
2014-04-18 10:50:46.000081 [30528] End of zbx_vc_add_value():SUCCEED
....
2014-04-18 10:50:46.000426 [30528] query [txnlev:1] [insert into history (itemid,clock,ns,value) values .... (63512,1397803833,121761075,0.000000) ....
....
2014-04-18 10:50:46.000461 [30528] In zbx_vc_add_value() itemid:63512 value_type:0 timestamp:1397803833.121761075
2014-04-18 10:50:46.000461 [30528] End of zbx_vc_add_value():SUCCEED

Also, during the search of problem, i founded duplicatino of items in db:

mysql> select * from history where itemid="41792";
|  41792 | 1397799152 | 0.0000 | 549560335 | 
|  41792 | 1397799152 | 0.0000 | 549560335 | 
|  41792 | 1397799272 | 0.4000 |  15553827 | 
|  41792 | 1397799273 | 0.4000 |  15553827 | 
|  41792 | 1397799392 | 0.2000 | 517255232 | 
|  41792 | 1397799392 | 0.2000 | 517255232 | 
|  41792 | 1397799512 | 0.0000 |   4664127 | 
|  41792 | 1397799512 | 0.0000 |   4664127 | 
|  41792 | 1397799512 | 0.0000 |   4664127 | 
|  41792 | 1397799632 | 0.1000 | 740457948 | 
|  41792 | 1397799632 | 0.1000 | 740457948 | 
|  41792 | 1397799752 | 0.1000 |  82475922 | 
|  41792 | 1397799752 | 0.1000 |  82475922 | 
|  41792 | 1397799872 | 0.1000 | 182039807 | 
|  41792 | 1397799872 | 0.1000 | 182039807 | 
|  41792 | 1397799992 | 0.4000 | 143365840 | 
|  41792 | 1397799992 | 0.4000 | 143365840 | 
|  41792 | 1397799992 | 0.4000 | 143365840 | 
|  41792 | 1397800112 | 0.0000 | 784272379 | 
|  41792 | 1397800112 | 0.0000 | 784272379 |


 Comments   
Comment by Oleksii Zagorskyi [ 2014 Apr 19 ]

Duplicated active agents in network ?
just identical time stamp is strange ...

Or DB specific configuration/issue

Comment by DKoshelev [ 2014 Apr 19 ]

Agents is passive, proxy is passive. There is only one agent on this host. DB configured from default screens.
After some tests we figured out that this is due to restart of zabbix server. After restart server the count of "processed values" in zabbix performance screen is doubled. Some times it doubles, sometime not. When it is not doubled, there is no incorrect values. We will try to debug it closer.

Comment by DKoshelev [ 2014 May 23 ]

This problems was due to low performance. After upgrading hardware on zabbix server all looks fine.

Comment by Oleksii Zagorskyi [ 2014 May 23 ]

Closing then.

Comment by Aleksandrs Saveljevs [ 2014 May 23 ]

Zabbix should not exhibit buggy behavior if run on an old hardware. Reopening...

Comment by Glebs Ivanovskis (Inactive) [ 2017 Oct 24 ]

Given the name of item in question ("... RPS" aka "requests per second") I can assume that it was stored as "Change" or "Change per second". And given that proxy was passive I claim that this was fixed in ZBX-12281.

Closing as Duplicate.

Generated at Fri Apr 26 06:51:36 EEST 2024 using Jira 9.12.4#9120004-sha1:625303b708afdb767e17cb2838290c41888e9ff0.