[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]

Re: [K12OSN] ldap problems...HELP!!!



Tait, 
 At my real job we used to use lots of power vaults.  Personally I've
lost lots of data because of them and I do not trust them much these
days.  Several times data was lost when drives failed and the hot spare
tried to kick in.  During the time between the failure and the hot spare
spinning up my Oracle rollback segments and archive logs were corrupted
trashing a 250gb database.  Believe it or not this happened twice in the
span of 2 months.  The first time it took 2 days to recover from a
backup.  The second time it took a few hours to roll in the archive
logs.  Not very good.  This January we moved to Sun and SAN disk and
have been sailing since.  Lots of the other wintel guys including the
exchange servers were moved from Power Vaults to SAN disk because of
similar problems they have.  Reguarly they would loose a drive and then
loose the hot spare as it tried to take over.  It happened twice in a
week on two separate servers.  Even before we used Dell our IBM EXP15s
were not much better.  With them the battery would go dead in the array
due to some kind of defect.  When you had a hard power outage you simply
lost the configuration data and everything on the array.  On the other
hand I've been using a Sun A1000 and D1000 Disk array for 3 years now
and have never had any problems except for the battery going dead on one
but at least it told us!  I'd blame it on the powervault and try another
fresh one.

Jack

On Thu, 2003-07-31 at 00:06, Tait Shrum wrote:
> I thought I would add somethings to this thread - I work with Barry at
> Bryant and have had trouble with this server for some time and finally
> gave up.
> 
> It is our only Redhat 9 server at this site.  We have 4 other RedHat 9
> servers, but they aren't causing any problems.  When I initially began to
> have issues with this server, I decided to "try" to eliminate network
> errors.  I turned off bonding, I hooked the nic up to a port directly on
> our core gigabit switch rather than on the gigabit switches in the Dell
> cabinet the servers are stored in.
> 
> I looked at the switches to see if we were receiving any funky stuff on
> the port the server was hooked up to, and there wasn't anything there.
> 
> A couple nights ago, I was eating dinner and thought about problems we
> were having with GradeQuick last year.  It would drop connections for
> certain people causing data corruption in their grade book files they had
> open.  We basically chalked it up to "weirdness" and restored stuff from
> the backup.  Those problems happened right after we setup that new server.
> 
> So - to tie all this together, it has six 146 GB drives in the main
> server.  Connected to it via the PERC 3 card is a Dell PowerVault that has
> 13 or 14 146 GB drives in it.  Do you all think the PowerVault is giving
> us fits?  That is where all the data users data "lives" on the server.
> 
> I'm just brain-storming hoping that others might have a PowerVault and
> could share some information about it.  We always appreciate the help we
> get from this list!  It's proven to be very valuable to us!
> 
> Tait Shrum
> 
> > Barry....keep us posted on how yoou make out...I'm watching this thread
> > with interest.  I haven't run into this yet, but would like to be prepared
> > should I come across it....Let us know if and how you fix it.
> >
> > David Trask
> >
> > k12osn redhat com writes:
> >>I need to make sense of some problems we are having...
> >>any help narrowing down our problems would be greatly appreciated at
> >>this point...
> >>I am about out of ideas.
> >>
> >>let me know if I should post my config files...
> >>
> >>We implemented ldap last year, and the transition went very smooth.
> >>we were on samba 2.2.8a, and openldap 2.1.17
> >>so, samba 2.2 schema, and we used the idealix smbldap-tools.
> >>RedHat 9
> >>
> >>I use a windows client ldap browser, to check how our ldap was doing, and
> >>we were using ldbm backend, with cachesize 1000, and dbcachesize
> >>1000000, and it worked great.
> >>the softerra ldap browser would pop up our 1000 users instantly, and I
> >>could click on any of them and get instant results..
> >>
> >>we had a server crash, and had to start over from scratch....so, we
> >>decided to get samba 3.0 working, in compat mode with
> >>our samba 2.2 schema.  I went ahead and used the schema file from samba
> >>3.0, and modified it, so our ldap entries still worked
> >>2.2 style.  We still use the idealix smb-tools to add, and delete
> >> entries.
> >>
> >>The problem is, it works, but we are seeing some erratic behavior out of
> >>the server now....
> >>We have tried, openldap 2.0.27, 2.1.17, 2.1.22, and nothing has fixed
> >>the strange things we are seeing.
> >>
> >>strange problems:
> >>1) the softerra ldab browser now takes forever to open and view the ldap
> >>directory
> >>we need to determine why.....softerra comes with openldap.org setup as
> >>another ldap directory,
> >>and their entries pop up instantly....what could we possibly have
> >>configured wrong ?
> >>is this limited to remote queries, or does it affect localhost
> >>also...slapcat works instantly...
> >>no delay there...samba log-ons seem just a tad slower than they did last
> >>year, but then again,
> >>we were very impressed at samba-ldap speed of logins compared to what we
> >>had seen with
> >>samba, and smbpasswd auth
> >>
> >>2) since starting over from scratch on the server, re-loading redhat
> >>9...our putty terminals
> >>will drop our shell connections several times during the day....it is
> >>very frustrating, and we have no idea what is causing it.
> >>when it happens, sometimes we can't log right back in, nad have to wait
> >>a sec...the strange thing is that this server is the only one doing it...
> >>all 6 other linux servers have no problems..
> >>thinking it may be hardware related, we have tried disabling the onboard
> >>gigabit nics in our server, and installing 2 new
> >>gigabit nic's, and we are still seeing the same thing...I have also
> >>wondered if it was ldap related, and that ldap was runnign too slow to
> >>keep up
> >>and everynow and then , connections would terminate
> >>
> >>3)openldap when we moved to 2.1.21 rpms, and then again on 2.1.22 rpms,
> >>gave us an error when database backend was set to ldbm.
> >>something about not finding the module, but then the module it was aking
> >>for was not even on the system.  i looked at the source rpm,
> >>and the spec file shows --enable-ldbm is there, and nothing about
> >>compiling ldbm as a module is in there...so I have no idea about that
> >> one.
> >>we had to specify bdb for those rpms, and the ldap server acted very
> >>funny while that was set...timing out on entries, and searches, and
> >>smbldap-usermod.pl commands....causing database corruption.
> >>I have since reverted back to apt-getting  openldap 2.0.27 that came
> >>with redhat 9.and it is working , but still having strange problems...
> >>
> >>so, since I have ldbm working, with 2.0.27, the only thing different
> >>from a really great system last year, and this quirky one runnign right
> >>now, is the
> >>samba 2.28a, and samba 2.2 ldap schema...
> >>
> >>I am going to revert back to 2.2.8a tommorrow and see if it works as
> >>well as it used to, but if anyone can give me more input, I would
> >>appreciate it.
> >
> >
> > David N. Trask
> > Technology Teacher/Coordinator
> > Vassalboro Community School
> > dtrask vcs u52 k12 me us
> > (207)923-3100
> >
> >
> > _______________________________________________
> > K12OSN mailing list
> > K12OSN redhat com
> > https://www.redhat.com/mailman/listinfo/k12osn
> > For more info see <http://www.k12os.org>
> >
> 
> 
> _______________________________________________
> K12OSN mailing list
> K12OSN redhat com
> https://www.redhat.com/mailman/listinfo/k12osn
> For more info see <http://www.k12os.org>




[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]