[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [leafnode-list] leafnode went tits-up. CPR ineffective.
Dnia Thu, 13 Mar 2003 00:41:14 +0100 niejaki(a) Matthias Andree
<ma@xxxxxxxxxxxxxxxxxxxxxxxxxxxx> napisał(a):
> Try 1.9.36.rel first, it has received fixes to signal handling that
> may help you already. Do not use 1.9.33 to 1.9.35.
ok, shall do that ASAP.
> Second, if 1.9.36 still shows this behaviour, and you know fetchnews
> has finished downloading the newsgroups (modem lights/NIC lights
> remain dark or something), please try to obtain a stack backtrace at
> the time you know it hangs.
You, know its funny. It's not a hang or crash actually. it is different.
after 3-4 hours it seems fetchnews starts actually fetching. one run of
fetchnews takes circa 17 hours (left fetchnews -vvvv on a console, and
actually right now, in this very moment, it ends the fetching...). for
each server it took that 3-4 hoursa...
> > Newsgroup name conflict: za.misc vs. za.misc
> > Newsgroup name conflict, chose za.misc
>
> Hum, strange. What does "locale" print for the root user?
root@sandra:/var/spool/news# locale
LANG=POSIX
LC_CTYPE="POSIX"
LC_NUMERIC="POSIX"
LC_TIME="POSIX"
LC_COLLATE="POSIX"
LC_MONETARY="POSIX"
LC_MESSAGES="POSIX"
LC_PAPER="POSIX"
LC_NAME="POSIX"
LC_ADDRESS="POSIX"
LC_TELEPHONE="POSIX"
LC_MEASUREMENT="POSIX"
LC_IDENTIFICATION="POSIX"
LC_ALL=POSIX
hmmm... wait a moment, LANG and LC_ALL are without the quotes... (?)
I'll check that.
> Has it died? Or has it finished? If you have a quick machine (some
> hundred MHz) with enough RAM to cache the articles it has just
> downloaded, then the child may read the articles from the cache rather
> than from disk and finish in fractions of a second, faster than you
> type"ps -A".
impossible. there's 512MB ram, but for cache/buffers usually up to 100MB
is available (its quite a handful there - 15GB squid-cache (takes 300MB
easy), mail, news, www, snmp, smb (fileserver), telnet/ssh), two network
game servers (halflife, and quake, afair). the main disk is a 40GB
maxtor 5200rpm, usually it chirping every second or so. also its just a
celeron 333mhz.
> > I've tried looking at leafnodes group files, came out zilcho.
>
> Where exactly have you looked?
www.google.com
groups.google.com
http://www.leafnode.org/archive/index.html
> What have you found, and what was
> different from what you had expected?
I've read, that removing /var/spool/news/leaf.node/news.supermedia.pl,
etc, could lead to something. or other files there. nope, not that
using fetchnews -vvvv to get as much data as possible
check if there is any reading, by vmstat, lsof, etc. usually yes, but
its a mother of job. although, yes, I found, that fetchnews hadn't wrote
anything to disk. at least not in 15 minutes from start...
--
Witold Władysław Wojciech Wilk [ rwt: +48(0)322270471 ] (ICQ#65665716)
[ n7110: +48(0)605066384 ] <maniack@xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx>
< GG#3211630 > { GIT d- s+:- a--- C++ UL++++ P+ L+++ !E W N+++ o? K? y++
O? !M V? PS+ PE+++ Y+ PGP !t 5? !X R+ !tv b++++ DI? D+++ G e- !h r- !w }
--
leafnode-list@xxxxxxxxxxxxxxxxxxxxxxxxxxxx -- mailing list for leafnode
To unsubscribe, send mail with "unsubscribe" in the subject to the list