Upgrade SLES8 > 9 with ora 2007-01-26 - By Alexei_Roudnev
Back
--- Off topic ---
O not. All our servers works without ANY glitch in production (except caused
by old hardware), including such fragile things as backups, which can run
few weeks or even month without my attention (except tapes rotation). But
when testing in the lab... you can't image a number of problems and glitches
which I saw in SLES + (Oracle) + ( cluster) + (iSCSI) + (OCFSv2) combination
/side effect - I plan to use OCFSv2 in production - for application source
FS< but not for _customers data FS_. You can guess, why...
(Last glitch, little aside - Linux NFS client is not compatible with
'Windows for Unix 3.5 NFS', IF linux is configured with fully qualified
host name.
Discovered during Oracle Linux tests. I get mad when I find it, after a few
hours of checking and comparing everything)
I am too lazy to wake up weekly at night, and we have a very small team (2
for everything including design and some development, on the whole data
center, including network, SAN, backups, Linuxes, Solarises, and so on - the
whole zoo), so I prefer to over-estimate problems. Anyway, we was one of the
first to use iSCSI and use FCP on NetApps, and did a risky decision to
select prefer SLES to RHEL (as a result, I test Oracle Linux in the lab
again).
> Don't be so put off by what Alexei is saying, he's the Greek tragedy
character
> in our lists :-) He probably walks in the datacenter wailing "oh Gods, my
> beloved servers are condemned, doomed they are, oh Lords".
You are correct that I (sometimes) am walking in my Linux lab and cry _who
did it!@#?_ But, to be perfect, I rarely could say it about SLES - more
often about our own developers (I have 2 my favorite companies to say it
aboththeir products... will not disclosure them here). Last case was about
_why simple, primitive multipath don't work, in a very standard
configuration_? (I saw the same with QLogic cards some time ago, but wrote
it out to the ancient QLA cards... now I can see, that it is something about
FCP switch and FCP drivers or even SCSI layer.)
(On the other hand, when we got new internal build of our new product, I
find a few bugs in 10 minutes - so, may be, Linux is just my victim).
To be serious, I dream about RedHat stability, integrated with SLES
management leyer and SLES set of packages. And about REAL testing by the
vendor(s).
--
To unsubscribe, email: suse-oracle-unsubscribe@(protected)
For additional commands, email: suse-oracle-help@(protected)
Please see http://www.suse.com/oracle/ before posting
|
|
