cloudstack-users mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Andrija Panic <andrija.pa...@gmail.com>
Subject Re: Strange bug? "spam" in management log files...
Date Fri, 05 Jun 2015 14:24:05 GMT
Hi,

any hint on how to proceed ?

on haproxy I see rougly 50%/50% sessions across 2 backend servers.
 But inside DB, it all points to the one mgmt_server_ip...

Thanks,
Andrija

On 4 June 2015 at 19:27, Andrija Panic <andrija.panic@gmail.com> wrote:

> And if of any help another hint:
>
> while Im having this lines sent to logs in high volume...if I stop second
> mgmt server, first one (that is making all these lines, doesnt stop to make
> them), so log is still heavily writen to - only when I also restart mgmt on
> 1st node (2nd node is down), then these log lines dissapear.
>
> Thx
>
> On 4 June 2015 at 19:19, Andrija Panic <andrija.panic@gmail.com> wrote:
>
>> And I could add - these lines (in this volume) only appears on first mgmt
>> server (Actually I have 2 separate, but identical ACS installations, and
>> same behaviour).
>>
>> On 4 June 2015 at 19:18, Andrija Panic <andrija.panic@gmail.com> wrote:
>>
>>> Just checked, in the HOSTS table, all agents are connected (via haproxy)
>>> to the first mgmt server...I just restarted haproxy, and still inside the
>>> DB, it says same mgmt_server_id for all agents - which is not really true.
>>>
>>> Actually, on the haproxy itslef (statistics page) I can see almoust
>>> 50%-50% distribution across 2 backends - which means by haproxy it should
>>> be fine.
>>> total 18 agents, 10 goes to 1 backend, 8 goes to other backend (ACS mgmt
>>> server)
>>>
>>> This is our haproxy config, I think it's fine, but... DB says
>>> differently, althouh haproxy statistick say all fine
>>>
>>> ### ACS 8250
>>> #######################################################################################
>>> frontend front_ACS_8250 10.20.10.100:8250
>>>         option tcplog
>>>         mode tcp
>>>         default_backend back_8250
>>> backend back_8250
>>>         mode tcp
>>>         balance source
>>>         server acs1_8250 10.20.10.7:8250 check port 8250 inter 2000
>>> rise 3 fall 3
>>>         server acs2_8250 10.20.10.8:8250 check port 8250 inter 2000
>>> rise 3 fall 3
>>>
>>> ##################################################################################################
>>>
>>> Any info on how to proceed with this, since because of these lines, it
>>> makes mgmt logs almoust unreadable... :(
>>>
>>> Thanks,
>>> Andrija
>>>
>>> On 4 June 2015 at 19:00, Andrija Panic <andrija.panic@gmail.com> wrote:
>>>
>>>> Thanks Koushik,
>>>>
>>>> I will check and let you know - but 11GB log file for 10h ?  I dont
>>>> expect this is expected :)
>>>> I understand that the message is there because of setup, just an awful
>>>> lot of lines....
>>>>
>>>> Will check thx for the help !
>>>>
>>>> Andrija
>>>>
>>>> On 4 June 2015 at 18:53, Koushik Das <koushik.das@citrix.com> wrote:
>>>>
>>>>> This is expected in a clustered MS setup. What is the distribution of
>>>>> HV hosts across these MS (check host table in db for MS id)? MS owning
the
>>>>> HV host processes all commands for that host.
>>>>> Grep for the sequence numbers (for e.g. 73-7374644389819187201) in
>>>>> both MS logs to correlate.
>>>>>
>>>>>
>>>>>
>>>>> On 04-Jun-2015, at 8:30 PM, Andrija Panic <andrija.panic@gmail.com>
>>>>> wrote:
>>>>>
>>>>> > Hi,
>>>>> >
>>>>> > I have 2 ACS MGMT servers, loadbalanced properly (AFAIK), and
>>>>> sometimes it
>>>>> > happens that on the first node, we have extremem number of folowing
>>>>> line
>>>>> > entries in the log fie, which causes many GB log in just few hours
>>>>> or less:
>>>>> > (as you can see here they are not even that frequent, but sometimes,
>>>>> it
>>>>> > gets really crazy with the speed/numer logged per seconds:
>>>>> >
>>>>> > 2015-06-04 16:55:04,089 DEBUG [c.c.a.m.ClusteredAgentManagerImpl]
>>>>> > (AgentManager-Handler-29:null) Seq 73-7374644389819187201: MgmtId
>>>>> > 90520745449919: Resp: Routing to peer
>>>>> > 2015-06-04 16:55:04,129 DEBUG [c.c.a.m.ClusteredAgentManagerImpl]
>>>>> > (AgentManager-Handler-28:null) Seq 1-3297479352165335041: MgmtId
>>>>> > 90520745449919: Resp: Routing to peer
>>>>> > 2015-06-04 16:55:04,129 DEBUG [c.c.a.m.ClusteredAgentManagerImpl]
>>>>> > (AgentManager-Handler-8:null) Seq 73-7374644389819187201: MgmtId
>>>>> > 90520745449919: Resp: Routing to peer
>>>>> > 2015-06-04 16:55:04,169 DEBUG [c.c.a.m.ClusteredAgentManagerImpl]
>>>>> > (AgentManager-Handler-26:null) Seq 1-3297479352165335041: MgmtId
>>>>> > 90520745449919: Resp: Routing to peer
>>>>> > 2015-06-04 16:55:04,169 DEBUG [c.c.a.m.ClusteredAgentManagerImpl]
>>>>> > (AgentManager-Handler-30:null) Seq 73-7374644389819187201: MgmtId
>>>>> > 90520745449919: Resp: Routing to peer
>>>>> > 2015-06-04 16:55:04,209 DEBUG [c.c.a.m.ClusteredAgentManagerImpl]
>>>>> > (AgentManager-Handler-27:null) Seq 1-3297479352165335041: MgmtId
>>>>> > 90520745449919: Resp: Routing to peer
>>>>> > 2015-06-04 16:55:04,209 DEBUG [c.c.a.m.ClusteredAgentManagerImpl]
>>>>> > (AgentManager-Handler-2:null) Seq 73-7374644389819187201: MgmtId
>>>>> > 90520745449919: Resp: Routing to peer
>>>>> > 2015-06-04 16:55:04,249 DEBUG [c.c.a.m.ClusteredAgentManagerImpl]
>>>>> > (AgentManager-Handler-4:null) Seq 1-3297479352165335041: MgmtId
>>>>> > 90520745449919: Resp: Routing to peer
>>>>> > 2015-06-04 16:55:04,249 DEBUG [c.c.a.m.ClusteredAgentManagerImpl]
>>>>> > (AgentManager-Handler-7:null) Seq 73-7374644389819187201: MgmtId
>>>>> > 90520745449919: Resp: Routing to peer
>>>>> > 2015-06-04 16:55:04,289 DEBUG [c.c.a.m.ClusteredAgentManagerImpl]
>>>>> > (AgentManager-Handler-3:null) Seq 1-3297479352165335041: MgmtId
>>>>> > 90520745449919: Resp: Routing to peer
>>>>> > 2015-06-04 16:55:04,289 DEBUG [c.c.a.m.ClusteredAgentManagerImpl]
>>>>> > (AgentManager-Handler-5:null) Seq 73-7374644389819187201: MgmtId
>>>>> > 90520745449919: Resp: Routing to peer
>>>>> > 2015-06-04 16:55:04,329 DEBUG [c.c.a.m.ClusteredAgentManagerImpl]
>>>>> > (AgentManager-Handler-1:null) Seq 1-3297479352165335041: MgmtId
>>>>> > 90520745449919: Resp: Routing to peer
>>>>> > 2015-06-04 16:55:04,330 DEBUG [c.c.a.m.ClusteredAgentManagerImpl]
>>>>> > (AgentManager-Handler-15:null) Seq 73-7374644389819187201: MgmtId
>>>>> > 90520745449919: Resp: Routing to peer
>>>>> > 2015-06-04 16:55:04,369 DEBUG [c.c.a.m.ClusteredAgentManagerImpl]
>>>>> > (AgentManager-Handler-11:null) Seq 1-3297479352165335041: MgmtId
>>>>> > 90520745449919: Resp: Routing to peer
>>>>> > 2015-06-04 16:55:04,369 DEBUG [c.c.a.m.ClusteredAgentManagerImpl]
>>>>> > (AgentManager-Handler-17:null) Seq 73-7374644389819187201: MgmtId
>>>>> > 90520745449919: Resp: Routing to peer
>>>>> > 2015-06-04 16:55:04,409 DEBUG [c.c.a.m.ClusteredAgentManagerImpl]
>>>>> > (AgentManager-Handler-14:null) Seq 1-3297479352165335041: MgmtId
>>>>> > 90520745449919: Resp: Routing to peer
>>>>> > 2015-06-04 16:55:04,409 DEBUG [c.c.a.m.ClusteredAgentManagerImpl]
>>>>> > (AgentManager-Handler-12:null) Seq 73-7374644389819187201: MgmtId
>>>>> > 90520745449919: Resp: Routing to peer
>>>>> >
>>>>> >
>>>>> > We have haproxy VIP, to which SSVM connects, and all cloudstack
>>>>> agents
>>>>> > (agent.properties file).
>>>>> >
>>>>> > Any suggestions, how to avoid this - I noticed when I turn off
>>>>> second ACS
>>>>> > MGMT server, and then reboot first one (restart
>>>>> cloudstack-management) it
>>>>> > stops and behaves nice :)
>>>>> >
>>>>> > This is ACS 4.5.1, Ubuntu 14.04 for mgmt nodes.
>>>>> >
>>>>> > Thanks,
>>>>> > --
>>>>> >
>>>>> > Andrija Panić
>>>>>
>>>>>
>>>>
>>>>
>>>> --
>>>>
>>>> Andrija Panić
>>>>
>>>
>>>
>>>
>>> --
>>>
>>> Andrija Panić
>>>
>>
>>
>>
>> --
>>
>> Andrija Panić
>>
>
>
>
> --
>
> Andrija Panić
>



-- 

Andrija Panić

Mime
  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message