[Users] how to use sample control for port counters

Wed Jan 13 09:29:11 PST 2016

12.01.2016 20:40, Black.S пишет:
> 11.01.2016 23:59, Coulter, Susan K пишет:
>>
>> The perfquery command will tell you about the throughput and errors 
>> on a port.
>> The smpquery command will tell you more of what you might want to 
>> know about a port.
>> They come from the infiniband-diags module / RPM.
>> Examples are below.
>>
>> Another option is to use the PerfMgr which is part of the OpenSM code 
>> base, if you are running osm on a host.
>> There is an example of that below too, using the default port/localhost.
>>
>> [root at mu-master ~]# perfquery 199 1
>> # Port counters: Lid 199 port 1 (CapMask: 0x1400)
>> PortSelect:......................1
>> CounterSelect:...................0x0000
>> SymbolErrorCounter:..............0
>> LinkErrorRecoveryCounter:........0
>> LinkDownedCounter:...............0
>> PortRcvErrors:...................0
>> PortRcvRemotePhysicalErrors:.....0
>> PortRcvSwitchRelayErrors:........0
>> PortXmitDiscards:................0
>> PortXmitConstraintErrors:........0
>> PortRcvConstraintErrors:.........0
>> CounterSelect2:..................0x00
>> LocalLinkIntegrityErrors:........0
>> ExcessiveBufferOverrunErrors:....0
>> VL15Dropped:.....................0
>> PortXmitData:....................4294967295
>> PortRcvData:.....................4294967295
>> PortXmitPkts:....................4294967295
>> PortRcvPkts:.....................4294967295
>> PortXmitWait:....................2503387680
>>
>>
>> [root at mu-master ~]# smpquery pi -L 199
>> # Port info: Lid 199 port 0
>> Mkey:............................<not displayed>
>> GidPrefix:.......................0xfe80000000000000
>> Lid:.............................199
>> SMLid:...........................250
>> CapMask:.........................0x2510868
>> IsTrapSupported
>> IsAutomaticMigrationSupported
>> IsSLMappingSupported
>> IsSystemImageGUIDsupported
>> IsCommunicatonManagementSupported
>> IsVendorClassSupported
>> IsCapabilityMaskNoticeSupported
>> IsClientRegistrationSupported
>> DiagCode:........................0x0000
>> MkeyLeasePeriod:.................0
>> LocalPort:.......................1
>> LinkWidthEnabled:................1X or 4X
>> LinkWidthSupported:..............1X or 4X
>> LinkWidthActive:.................4X
>> LinkSpeedSupported:..............2.5 Gbps or 5.0 Gbps or 10.0 Gbps
>> LinkState:.......................Active
>> PhysLinkState:...................LinkUp
>> LinkDownDefState:................Polling
>> ProtectBits:.....................2
>> LMC:.............................0
>> LinkSpeedActive:.................10.0 Gbps
>> LinkSpeedEnabled:................2.5 Gbps or 5.0 Gbps or 10.0 Gbps
>> NeighborMTU:.....................4096
>> SMSL:............................0
>> VLCap:...........................VL0-3
>> InitType:........................0x00
>> VLHighLimit:.....................0
>> VLArbHighCap:....................8
>> VLArbLowCap:.....................8
>> InitReply:.......................0x00
>> MtuCap:..........................4096
>> VLStallCount:....................0
>> HoqLife:.........................31
>> OperVLs:.........................VL0-3
>> PartEnforceInb:..................0
>> PartEnforceOutb:.................0
>> FilterRawInb:....................0
>> FilterRawOutb:...................0
>> MkeyViolations:..................0
>> PkeyViolations:..................0
>> QkeyViolations:..................0
>> GuidCap:.........................128
>> ClientReregister:................0
>> McastPkeyTrapSuppressionEnabled:.0
>> SubnetTimeout:...................18
>> RespTimeVal:.....................16
>> LocalPhysErr:....................8
>> OverrunErr:......................8
>> MaxCreditHint:...................0
>> RoundTrip:.......................0
>> CapabilityMask2:.................0x0000
>> LinkSpeedExtActive:..............No Extended Speed
>> LinkSpeedExtSupported:...........0
>> LinkSpeedExtEnabled:.............0
>>
>>
>> [root at mu-master ~]# telnet localhost 10000
>> Trying 127.0.0.1...
>> Connected to localhost.
>> Escape character is '^]'.
>> OpenSM $ ?
>> ? : Command not found
>>
>> Supported commands and syntax:
>> help [<command>]
>> quit (not valid in local mode; use ctl-c)
>> loglevel [<log-level>]
>> permodlog
>> priority [<sm-priority>]
>> resweep [heavy|light]
>> reroute
>> sweep [on|off]
>> status [loop]
>> logflush [on|off] -- toggle opensm.log file flushing
>> querylid lid -- print internal information about the lid specified
>> portstatus [ca|switch|router]
>> switchbalance [verbose] [guid]
>> lidbalance [switchguid]
>> dump_conf
>> update_desc
>> version -- print the OSM version
>> perfmgr(pm) [enable|disable
>> |clear_counters|dump_counters|print_counters(pc)|print_errors(pe)
>> |set_rm_nodes|clear_rm_nodes|clear_inactive
>> |set_query_cpi|clear_query_cpi
>> |dump_redir|clear_redir
>> |sweep|sweep_time[seconds]]
>> dump_portguid [file filename] regexp1 [regexp2 [regexp3 ...]] -- Dump 
>> port GUID matching a regexp
>>
>>
>>
>>> On Dec 26, 2015, at 12:40 PM, Black.S <Black.S52 at yandex.com> wrote:
>>>
>>>
>>>
>>> Hello all
>>>
>>> I want to monitoring ports in IB fabric.
>>>
>>> I happened to notice in the out of perfquery some counter like a ticks
>>> and port sampling. Its will be great for accurate monitoring. But I cant
>>> found some info about setup and using sampling controi for IB ports.
>>>
>>> How to configure and use sampling contol for IB ports? Can anybody give
>>> me documentations/links or examples? I cant found it In Google.
>>> Are there any restrictions on the amount of data collected in this
>>> sampling mode?
>>>
>>> If its offtop then redirect me please
>>>
>>> Thanks for you time
>>>
>>> _______________________________________________
>>> Users mailing list
>>> Users at lists.openfabrics.org <mailto:Users at lists.openfabrics.org>
>>> http://lists.openfabrics.org/mailman/listinfo/users
>>
>> ==================================================
>> Susan Coulter
>> HPC Network Technical Lead
>> (505) 667-8425
>> “Once in a while you get shown the light
>>     In the strangest of places if you look at it right” /Robert Hunter/
>> ==================================================
>>
>>
>>
>>
>>
> Hi  Susan Coulter
>
> Very thanks for your reply
>
> Very interesting and especially for perfmon. How I understand the 
> perfmon can dump all counters very fast (i hope it more fast then dump 
> it in cycle one by one). Thanks for this info.
>
> But nevertheless, can your or somebody explain how to use sampling 
> control mechanism for perfquery ?
> /"//perfqery //-c, --smplctl////show port samples control"/
>
> in this mode, I would like to collect all available counters of the IB 
> port with time stamp provided by "tick" counter from the device.
>
> The time stamp form tick counter will allows not depending on time 
> from host which collect counters from IB port .
>
> Sorry for my english
>

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openfabrics.org/pipermail/users/attachments/20160113/c36fe392/attachment.html>