query FP_QUERY_RESULT_CODE_ABORT

Question

Hi, When I tried to do query against a pool using the api, I got the error: FP_QUERY_RESULT_CODE_ABORT. I then used the JCASScript to do similar query ( just use the 'query' command  with default 'start' and 'end' time settings ), I got: Number of clips returned:     0All results returned:         FalsePoint of incompletion:        immediate And when I looked at the logfile, I saw some exceptions:        [EXCEPTION]     In 'FPSmartPacket.cpp' at line 253: Exception error=-10152 syserror=0 message=field id=113 missing trace=FPSmartPacket.setFetchCursor(113) Do you know what's the potential cause for this error? I have been trying to find the cause for a while without any success... I did similar steps against the public emc clusters and was able to fetch the query results ok. any comments are appreciated. Thanks in advance, Wil

wo2 · Answer

p.s., I was using JCASScript-Linux-GCC3.3-3.2.35.tgz

wo2 · Answer

Hi,

I noticed that for the cluster that the query failed with the exception and FP_QUERY_RESULT_CODE_ABORT status, the node in the cabinet has a software version of 2.3x, i.e.,

Does this means that the node is running with the Centrastar version 2.3.3-312-429-6618 and needs the Centera SDK 2.4?

it seems that there's some compatibility issue with the packet protocol between the 3.1/3.2 SDK and the 2.3.3-312-429-6618 os, and thus causing the missing field exception in

[EXCEPTION] In 'FPSmartPacket.cpp' at line 253: Exception

error=-10152
syserror=0
message=field id=113 missing
trace=FPSmartPacket.setFetchCursor(113)

As well, if I do need to get the old version of the 2.4 SDK, where can I get it? ( we don't own the Centera storage, and thus probbaly not be able to upgrade it to the latest version )

Thanks!

Wil

gstuartemc · Answer

The 10152 error can be ignored, it is purely down to additional fields that later versions of C* or the SDK use. It is not the cause of your query error. Similarly, you do not need to get a later version of the SDK.

Query Abort messages are normally caused by cluster issues e.g. too busy to process the query (which is a very intense operation in the older versions of C*), nodes down (which normally will give an INCOMPLETE error rather than an abort).

I would suggest that you raise a support query for this.

wo2 · Answer

Hi Graham,

Thanks a lot for the prompt response and info!

When you said 'raise the query support', did you mean contacting the EMC support? if so, is there any specific email or contact number that I can call? Sorry for my ignorance, it's my first time involved with a emc product, I tried to search for support in the site but got a few confusing results back...

In addition to contact the support, do you know is there any other things I can try to debug the issue? It seems that the error only happened with the 'clip enumeration' query command. The nodes health are good and online ( from the monitor discovery ), and I tried the clipExists command, it returned without any error with the correct answer. But with the Query command, it failed right the way... from the log, it seems that the sequences are:

FPPoolQuery_Open()

FPPoolQuery_FetchResults()

--- a few 10152 Exceptions...

FPPoolQuery_FetchResults()

FPPoolQuery_GetResultCode() -> 1 ( incomplete )

FPPoolQuery_FetchResults()

--- 10152 exceptions ---

FPPoolQuery_FetchResults()

FPPoolQuery_GetResultCode() -> 99 ( Progress )

FPPoolQuery_FetchResults()

--- 10152 exceptions ---

FPPoolQuery_FetchResults()

FPPoolQuery_GetResultCode() -> 4 ( ABORT )

FPPoolQuery_Close()

The FPPool_GetLastError() always return [0] though...

any comments are greatly appreciated...

Thanks again,

Wil

Centera

Was this post helpful?