| T.R | Title | User | Personal
Name | Date | Lines |
|---|
| 1079.1 | It could be a lot of reasons | TOOK::SHMUYLOVICH | | Fri May 31 1991 19:11 | 29 |
|
>I setup a export job and it was collecting ok for about 15 mins, suddenly,
>the background job went away.
>
>I checked the log file, and it shows ALERT-TERMREQ, thread termination
>requested. What happened to it? What can cause the job to be terminated.
ALERT-TERMREQ, thread termination requested means that writer subprocess
of this background process is dead. The writer subprocess is responsible for
writing data in RDB file and the only reason for its death I know is not enough
system quotas (I don't tell that there aren't others!).
Several questions:
1. What entity/entities did you export?
2. How many successful polls did they have?
If entity being exported is node (not node4!) could you, please, check if all
returned arguments are defined in the data dictionary? You can verify this by
doing Show for all partitions. If among others attributes you will something
like "no such entity" it means that in the response you have an extra attribute.
Sam S.
Historian/Exporter team
|
| 1079.2 | More details | ZPOVC::RAMARAJ | | Fri May 31 1991 22:11 | 28 |
| The exports were:
node4 zpoac5
node4 zpoac5 line qna-0
node4 zpoac5 circ qna-0
node4 zpoac8
node4 zpoac8 line sva-0
node4 zpoac8 circ sva-0
The export period is 1min for zpoac5 and 15 mins for zpoac8, and the target
is mccreports:mcc_export.rdb.
The successfull no. of polls varied from 5 to 50, because zpoac8 was setup
earlier for exporting.
No. of unsuccessful polls was 1 and 3, this was due to a system shutdown for
a few mins during the exporting period.
A show of all partitions did not give any error.
One strange thing is that it has only happened so far when I start exporting
from a node4 with QBUS(QNA-0). Had the same problem at a customer site too.
Could this be due the memory leak problem in exporting? What quotas can
be adjusted for better performance?
Raj
SWS Singapore
|
| 1079.3 | | TOOK::SHMUYLOVICH | | Mon Jun 03 1991 19:20 | 23 |
|
>Could this be due the memory leak problem in exporting? What quotas can
>be adjusted for better performance?
Please, check open file quota and Fillm.
I exported node4, node4 line and node4 circuit for more that a day
with an interval = 5 min and I did not found any memory leaks. There are
some leaks only when I exported node.
Note: you can watch for memory leaks ONLY after at least one successful
poll for each active exporting.
>The export period is 1min for zpoac5 and 15 mins for zpoac8, and the target
>is mccreports:mcc_export.rdb.
Period = 1 min is too short. As I remember in documentation we recommend
period >= 15 min. But it can not be reason for background to stop.
Sam
|
| 1079.4 | Need to know which quota | NSSG::R_SPENCE | Nets don't fail me now... | Wed Jun 05 1991 11:35 | 5 |
| Could someone please state the exact quota referred to be the previous
note as "open file quota"?
thanks
s/rob
|
| 1079.5 | FILLM | SWAM2::MCCARTHY_LA | Use an accordian, go to jail! | Wed Jun 05 1991 12:56 | 0 |
| 1079.6 | Exporter thread failed .... | EVTAI1::RENOUVEL | | Tue Aug 31 1993 11:45 | 13 |
| Could somebody tell me the way to know how to be sure
of the cause of the MCC_EXPORTER_BACKGROUND fail ...
I've read in the precedent notes that the UAF parameter FILLM was the
trouble.
I tried to put it at 512 but same problem.
What could I do please...
Patrick
Telecoms and Network France.
|