Poster: 17006 - Error reading object cache item already in the object id registry (4301593)

Poster: 17006 - Error reading object cache item already in the object id registry

The following errors are observed in event log when Post is stopped due to error:

Error    2013-12-13 01:01:27.783893 32253 4081792864 Poster: 17006 - Error reading object cache from file "/u03/splex/state/0x0a0300fb+PP+queue_name+sp_opst+o.SID2-o.SID2-objcache_sp_opst.28": Illegal state: Item 263290 already in the object id registry  (posting from SID1, queue queue_name, to SID2) [module osp]

Error    2013-12-13 01:01:27.784774 32253 4081792864 Poster: 15010 - Error reading pre-sync'd objcache for datasource 0x0a0300fb+PP+queue_name+sp_opst_mt+o.SID1-o.SID2, actid 28  (posting from RTLDB, queue queue_name, to RTLDB) [module opo]

The Post object cache for previous activation id is corrupted.

The easiest way to recreate it is to use the "copy cache" command in SharePlex. The command is issued on source. Here is the syntax:

Create the missing object cache on the target using the latest sp_ocap object cache from source and start post. Here are the steps involved:

1.Make a copy of the currently active config file, then activate it.

2.On source stop Capture and make sure that it is “stopped by user” and not “stopping”

3.Restart Capture

4. Issue “list config” to find the activation id. The # is denoted by the string after the word conf in the last line of the output. For example,  it is conf.30 in our example which tells us that the latest activation id is 30.

5.Go to vardir/state dir on the SOURCE and look for *objcache_sp_ocap.30 (which happens to be actid from 4)

6. Copy the above object cache file in binary format to vardir/state directory

7. On the target rename this file to same name as the name of the missing file reported in the target event_log, namely 0x0a0300fb+PP+queue_name+sp_opst+o.SID2-o.SID2-objcache_sp_opst.28

8. Make sure that the ownership/permissions of the newly created object cache are same as those of the other object cache files (identified by the string *objcache*) in the /vardir/state directory on target.

9. Start post.

--------------------------------------------------------------

ADDENDUM

With the advent of Open Targets, the object cache error messages have changed slightly to include two different components, namely the database identifier or name is preceded by ‘r.’ and not ‘o.’, and the process is denoted by sp_xpst and not sp_opst as the following error message shows:

sp_ctrl > show log reverse

Info     2018-11-30 15:46:52.192000 4592 7388 Poster exited with code=1, pid = 5

12  (posting from o.SID1, queue queue_name, to r.DATABASE_NAME)

Error    2018-11-30 15:46:51.114000 6060 8160 Poster stopped: Internal error encountered; cannot continue  (posting from o.SID1, queue queue_name, to r.DATABASE_NAME)

Error    2018-11-30 15:46:51.099000 6060 8160 Poster cannot read object cache for actid 5  (posting from o.SID1, queue queue_name, to r.DATABASE_NAME)

Notice   2018-11-30 15:46:51.099000 6060 8160 Poster: Input/output: File not found: D:/Program Files/Quest Software/SharePlex/vardir2100/state/0x0a3cb40a+PP+davldb0190+sp_xpst+o.SID1-r.DATABASE_NAME-objcache_sp_xpst.5  (posting from ldb0190+o.SID1, queue queue_name, to r.DATABASE_NAME) [module osp]

Info     2018-11-30 15:46:51.020000 6060 8160 Poster launched, pid = 6060  (posting from o.SID1, queue queue_name, to r.DATABASE_NAME)

The rest of the resolution remains same, namely recreating Post object cache on target from Capture object from source.

--------------------------------------------------------------------

I have tried my best to illustrate the steps as dictated by earlier reviewer. The steps are as illustrative as they can be and are comparable with other articles involving different object cache issues. Hope this helps. Covered the steps for copying the object cache but how about if there are DDL changes between activation? Post fails if we have a DDL mismatch and not safe to copy over cache between activation.