[ZBX-4270] IPMI insufficient resource Created: 2011 Oct 25  Updated: 2017 May 30  Resolved: 2012 Nov 01

Status: Closed
Project: ZABBIX BUGS AND ISSUES
Component/s: Server (S)
Affects Version/s: 1.8.8
Fix Version/s: None

Type: Incident report Priority: Trivial
Reporter: hamid sfandiari Assignee: Unassigned
Resolution: Duplicate Votes: 0
Labels: ipmi
Remaining Estimate: Not Specified
Time Spent: Not Specified
Original Estimate: Not Specified
Environment:

FreeBSD , ipmitool-1.8.11_2, openipmi-2.0.18_4, DL 360 G7 ilo v3.0

Attachments: PNG File IPMI item graph.png    
Issue Links:
Duplicate
duplicates ZBX-3243 Network error while retrieving IPMI data Closed

 Description   

I've set 3 items for my server to get sensor data each 30 seconds but it Frequently changed to not supported
zabbix_server.log:
8042:20111025:103124.814 Enabling IPMI host [zproxy01]
8282:20111025:103134.311 IPMI Host [zproxy01]: first network error, wait for 15 seconds
8281:20111025:103141.545 Disabling IPMI host [zproxy01]
8130:20111025:103225.023 Enabling IPMI host [zabbix]
8290:20111025:103242.032 IPMI Host [zabbix]: first network error, wait for 15 seconds
8295:20111025:103336.089 IPMI Host [zabbix]: first network error, wait for 15 seconds
8291:20111025:103342.064 Disabling IPMI host [zabbix]
8123:20111025:103456.984 Enabling IPMI host [zabbix]
8102:20111025:103501.839 Enabling IPMI host [zproxy01]
8287:20111025:103539.066 IPMI Host [zproxy01]: first network error, wait for 15 seconds
8289:20111025:103540.094 IPMI Host [zproxy01]: another network error, wait for 15 secon

when it changed to not supported i had check with manual "ipmitool -U zabbix -H 192.168.XX.XX -I lanplus -L ADMINISTRATOR sensor"
i've got the following message:
Error in open session response message : insufficient resources for session

Error: Unable to establish IPMI v2 / RMCP+ session

also the result of "ipmitool -U zabbix -H 192.168.XX.XX -I lanplus -L ADMINISTRATOR session info all" maybe usefull (at the next possible time)
session handle : 1
slot count : 4
active sessions : 60
user id : 2
privilege level : ADMINISTRATOR
session type : IPMIv2/RMCP+
channel number : 0x02
console ip : 192.168.XX.XX
console mac : 00:00:00:00:00:00
console port : 4040

session handle : 2
slot count : 4
active sessions : 60
user id : 2
privilege level : ADMINISTRATOR
session type : IPMIv2/RMCP+
channel number : 0x02
console ip : 192.168.XX.XX
console mac : 00:00:00:00:00:00
console port : 59029

session handle : 0
slot count : 4
active sessions : 60

session handle : 4
slot count : 4
active sessions : 60
user id : 2
privilege level : ADMINISTRATOR
session type : IPMIv2/RMCP+
channel number : 0x02
console ip : 192.168.XX.XX
console mac : 00:00:00:00:00:00
console port : 25072

 Comments   
Comment by richlv [ 2011 Oct 25 ]

could it be that you're just overloading the target ipmi device, thus starving it of the resources ?
if so, i don't think there's anything simple zabbix could help with

Comment by hamid sfandiari [ 2011 Oct 25 ]

do you have think that 3 item for an iLo by the interval of 60 sec caused to Overloading?
it's better that zabbix get all result of sensor items from iLo in one step and gathering them in local buffer then dispatching them as they needed (because of serial aspect of data gathering from IPMI)
each date fetching need authentication and it's time consuming step

i think that the following pesudo script could solve this issue and preventing iLo from overloading

interval 1
#> ipmitool IP .... sensor > buffer
#> grep "temp 02" buffer | zabbix_send $1
#> grep "temp 03" buffer | zabbix_send $1

interval 2
#> ipmitool IP .... sensor > buffer
#> grep "temp 02" buffer | zabbix_send $1
#> grep "temp 03" buffer | zabbix_send $1

Comment by richlv [ 2011 Oct 25 ]

do i understand it correctly that your suggestion is to merge requests to ipmi devices if they happen to be scheduled close enough ?

as for what could overload devices, hard to say. if ipmitool returns you an error, then that's what is happening
having said that, i have seen very serious issues with hp cli tools for querying disk & array status when calling them often enough. that was definitely one thing their testing did not try at all..

Comment by hamid sfandiari [ 2011 Oct 25 ]

yes, you understood correctly

but still i am thinking that closing session after querying (don't waiting for session timeout happen on ilo side) is more effective solution
by this explanation IPMI tool is not useful way for hp in ZABBIX
(sorry for my bad English )

Comment by Cristian Mammoli [ 2012 Mar 02 ]

Similar thing happens on Fujitsu Primergu Servers.

My env:
Zabbix 1.8.10
Server Fujitsu Primergy RX300 S6

Very often I get errors in Zabbix logs and when I query the serevr with ipmitool I get:

[root@srvzabbix ~]# ipmitool -H <IP>-U admin -P admin sensor get "Power Unit"
Locating sensor record...
Get Session Challenge command failed: Out of space
Error: Unable to establish LAN session
Get Device ID command failed
Unable to open SDR for reading
Sensor data record "Power Unit" not found!

Instead of:

[root@srvzabbix ~]# ipmitool -H <IP> -U admin -P admin sensor get "Power Unit"
Locating sensor record...
Sensor ID : Power Unit (0x41)
Entity ID : 19.0
Sensor Type (Discrete): Power Unit
States Asserted : Redundancy State
[Redundancy Lost]

Comment by Alexei Vladishev [ 2012 Aug 27 ]

I am not sure we can do anything with this issue. We are looking forward to optimizing IPMI and SNMP polling by using bulk requests. It may help once implemented.

Comment by richlv [ 2012 Nov 01 ]

should be solved by ZBX-3243

Generated at Tue May 06 07:27:14 EEST 2025 using Jira 9.12.4#9120004-sha1:625303b708afdb767e17cb2838290c41888e9ff0.