home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: comp.databases.ingres
- Path: sparky!uunet!elroy.jpl.nasa.gov!usc!sol.ctr.columbia.edu!eff!news.oc.com!spssig.spss.com!uchinews!gsbacd.uchicago.edu!cs_mj
- From: cs_mj@gsbacd.uchicago.edu (Mark Jaeger)
- Subject: Re: E_CLFE07 and E_SC022F and E_LQ002E
- Message-ID: <1992Sep3.165656.1@gsbacd.uchicago.edu>
- Lines: 75
- Sender: news@uchinews.uchicago.edu (News System)
- Organization:
- References: <1992Sep3.161408.8517@news.columbia.edu>
- Date: Thu, 3 Sep 1992 22:56:56 GMT
-
- In article <1992Sep3.161408.8517@news.columbia.edu>,
- alan@curta.cc.columbia.edu (Alan Crosswell) writes:
- > I keep seeing this errors under 6.3 and 6.4. Anybody know that they
- > really mean if anything?
- >
- > ::[4026 , 00846860]: Thu Sep 3 04:02:17 1992
- > E_CLFE07_BS_READ_
- > ERR Read from peer process failed; it may have exited.
- >
- > ::[4026 , 00846860]: Thu Sep 3 04:02:17 1992
- > E_SC022F_BAD_GCA_
- > READ SCF was unable to read a GCF block.
-
- I mostly know VMS, but maybe this will help. It looks like the front
- end went away, and the server tried to look for the next query from it,
- but couldn't. It doesn't look like cause for concern. In the last few
- months, this is the only similar thing we've seen:
-
- ::[II_DBMS_12D , 005FB140]: 6-AUG-1992 11:19:29
- E_GC0001_ASSOC_FAIL Association failure: partner abruptly released association
- ::[II_DBMS_12D , 005FB140]: 6-AUG-1992 11:19:29
- E_SC022F_BAD_GCA_READ SCF was unable to read a GCF block.
- ::[II_DBMS_12D , 005FB140]: 6-AUG-1992 11:19:29
- E_SC0123_SESSION_INITIATE Error initiating session.
- ::[II_DBMS_12D , 005FB140]: 6-AUG-1992 11:19:39
- E_GC0001_ASSOC_FAIL Association failure: partner abruptly released association
-
- Your E_CL error comes out of the compatibility library, which is like
- the server's kernel and provides a set of system calls (for server code)
- that are common across all platforms. The E_SC error comes out of the
- SCF (session control?), which is like the dbms's scheduler and
- coordinates all activities associated with a session. It says it
- couldn't read a GCF block, which is a communication block between the
- front end and the back end. It looks like the SCF tried to read a GCF
- block by making a read call to the CL. The CL couldn't satisfy the read
- because the front end process was gone. The CL logged the error and
- returned an error code to the SCF, which also logged an error.
-
- You might try to reproduce this by starting up a front end and then
- killing it in various ways: ^C, ^\, kill -9, kill -HUP, etc. See what
- gets logged. Try giving the interrupt in the middle of a select loop or
- cursor fetch loop, or after a cursor open but before the first fetch.
-
- > Also, for the first time ever on a SQL program that we run regularly, this
- > one came out:
- >
- > E_LQ002E _NODB The 'open ' query has been issued outside of a DBMS
- > session
- >
- > (and the saved query text was "xec..e ..e.y (c.f1) '19923BUSI6015B02'")
- >
- > However, there is no other useful indication in errlog to correlate this
- > to, other than those frequent E_CLFE07/E_SC022F errors.
-
- It looks like an "open cursor" statement was issued, but after the front
- end and back end lost contact somehow. Looks weird. In VMS, the front
- end and back end communicate over 4 different channels (I think). Each
- end gets two mailboxes to write to (for interprocess communication), one
- for queries/data and the other for status information. In your case, it
- looks like maybe the query/data channel got hosed, but the status
- channel was still there so the server could report this error message
- back to the front end.
-
- What is the saved query text after a normally executed "open"? What was
- the name of the cursor that you were opening, and what is its associated
- "select"?
-
- I'm shooting in the dark here, but maybe some of this will help.
-
- --Mark Jaeger internet: cs_mj@gsbvax.uchicago.edu
- Graduate School of Business yellnet: (312) 702-0328
- University of Chicago faxnet: (312) 702-0233
- Disclaimer: My opinions are my own and not those of my employer.
- Ich bin ein Virus. Mach' mit und kopiere mich in Deine .signature.
-
-