nelson@sun.soe.clarkson.edu (Russ Nelson) (07/22/89)
Sometimes relaynews fails with status 16, which is the value of ST_DISKFULL, which, of course, is an indication of a disk write error. It puts the batch into in.coming/bad. If I move the file(s) out of bad and then run newsrun, *every* one of the articles is a duplicate. In other words, a spurious error is being generated after all the articles have been processed. I say spurious because we are running with tens of megabytes free on our spool partition. Can anyone tell me how to find out *which* call to fulldisk is generating the error? I suppose that I could: #define fulldisk dump(__FILE__,__LINE__),fulldisk but that seems somehow dangerous. -- --russ (nelson@clutx [.bitnet | .clarkson.edu])|(70441.205@compuserve.com)
geoff@utstat.uucp (Geoff Collyer) (07/23/89)
With one exception (when the standard file descriptors are found to be closed during startup), relaynews writes an error message on standard error (a.k.a NEWSCTL/errlog) when it exits with non-zero status. In the case of a full disk, the name of the file being written when the disk filled is included in the error message. (Of course, the error message may not be visible if NEWSCTL's file system just filled.) Those of you running System V will see an apparently-spurious full-disk complaint when any file exceeds your ulimit; this most often hits the history and log files. -- Geoff Collyer utzoo!utstat!geoff, geoff@utstat.toronto.edu ``... skill such as yours is evidence of a misspent youth.'' - Herbert Spencer
karl@ddsw1.MCS.COM (Karl Denninger) (07/23/89)
In article <1989Jul22.211235.17447@utstat.uucp> geoff@utstat.uucp (Geoff Collyer) writes: >With one exception (when the standard file descriptors are found to be >closed during startup), relaynews writes an error message on standard >error (a.k.a NEWSCTL/errlog) when it exits with non-zero status. In the >case of a full disk, the name of the file being written when the disk >filled is included in the error message. (Of course, the error message >may not be visible if NEWSCTL's file system just filled.) Those of you >running System V will see an apparently-spurious full-disk complaint >when any file exceeds your ulimit; this most often hits the history and >log files. This is not the problem. We get the same spurious message on occasion here running Xenix System V/386 2.3.2. The batch IS correctly processed through to the end, but something which is done at the very end of the batch is returning a status 16... It is _not_ ulimit-related -- our ulimit is set to 32k! What is more strange is that it only happens once a day or so -- once out of about 20-30 batches on average! Nothing indicating the cause of the problem has been found in the log files. This is an annoyance, but not a serious problem -- we just "rm" the files in the "bad" area once in a while. It would, though, be nice to know what is doing this and squash it. -- Karl Denninger (karl@ddsw1.MCS.COM, <well-connected>!ddsw1!karl) Public Access Data Line: [+1 312 566-8911], Voice: [+1 312 566-8910] Macro Computer Solutions, Inc. "Quality Solutions at a Fair Price"
mann@intacc.uucp (Jeff Mann) (07/24/89)
I also experienced relaynews failing with status 16, even though it seemed that all the articles got installed. What was happening in fact was that I didn't have a junk group in active. Any batch with one or more article that didn't have valid newsgroups in my active would fail with status 16, and the batch would be placed in in.coming/bad. -- | Jeff Mann - Inter/Access, Toronto ...uunet!mnetor!intacc!mann | | "A picture is worth 256 thousand words" {utzoo, utgpu}!chp!intacc!mann |
kayvan@mrspoc.Transact.COM (Kayvan Sylvan) (07/25/89)
In article <1989Jul23.172948.19334@intacc.uucp> mann@intacc.uucp (Jeff Mann) writes: > I also experienced relaynews failing with status 16, even though it seemed > that all the articles got installed. What was happening in fact was that > I didn't have a junk group in active. Any batch with one or more article > that didn't have valid newsgroups in my active would fail with status 16, > and the batch would be placed in in.coming/bad. I am also experiencing the "relaynews fails with status 16" problem, but I *do* have a junk group in active. Does junk need to be known in my sys file? ---Kayvan -- Kayvan Sylvan @ Transact Software, Inc. -*- Mountain View, CA (415) 961-6112 Internet: kayvan@Transact.COM -*- UUCP: ...!{apple,pyramid,mips}!mrspoc!kayvan
nelson@sun.soe.clarkson.edu (Russ Nelson) (07/26/89)
In article <1989Jul22.211235.17447@utstat.uucp> geoff@utstat.uucp (Geoff Collyer) writes:
With one exception (when the standard file descriptors are found to be
closed during startup), relaynews writes an error message on standard
error (a.k.a NEWSCTL/errlog) when it exits with non-zero status.
Thanks, I didn't understand what I was seeing. At very least, I'm having
trouble running news2mail in a TCP/IP only environment, and it's returning
non-zero exit codes. I'll fix it and see if I still get bad batches.
--
--russ (nelson@clutx [.bitnet | .clarkson.edu])|(70441.205@compuserve.com)