Subject: Re: More info on SLES10 needed 2007-08-14 - By Alexei_Roudnev
Back
Ok.
1) SLES10 SP1 - when running Oracle data guard in SYNC mode, lost log
synchronization, message is 'async io error' (on network I think). Never
happen with SLES9 or RHEL4.
2) SLES10 SP1 - if using iSCSI, then no any way to mount file systems
(_netdev not implemented but 'hotplug' don't work with multipath);
3) New Open-iSCSI means _no multiport support, need to run multipath_;
poorly documented;
4) (again) iSCSI is not aliened with LVM and LVM is not aligned with
multipath. So you must configure everything manually.
5) Network Update - no normal (by yast) way to configure update server.
6) Installation - when few installation sources are configured, yast2
install menu starts about 5 - 10 minutes because it reread all sources.
Getting source to DOWN state do not help.
7) No support for older thread library (it is temporary problem, of course,
because this library is not recommended).
8) You can't use LVM with yast if you create physical volumes on the whole
devices (not on the partitions).
9) Heartbeat2 - SLES9 had full support for heartbeat in yast2; SLES10 have
only a very primitive support there.
10) No standard way to control duplex setting on the interfaces (but it's a
problem of all SuSe-s).
Of course, working with RHEL is all hassle:
- bad rpm naming cause a problem in determining _what is really installed_;
- no integrated installation and ad,ministration system, makes support
really guru-s task;
- software set is poor by default;
- no AppArmor (SELinix is not equivalent replacement).
- numerous 'druids', 'helpers', 'configers' makes administrative tasks a
really adventure (and most people roll back to 'vi' when working with RHEL).
But it works. LVM is aligned with multipath, and both with iSCSI, and fstab
supports _netdev. No any problems with DataGuard streams; no any problems
with LVM, no any problems with async io, no forgotten 'reboot on panic'
settings... Looks as old, iron made, very primitive but reliable car.
I hoped that it make space for better systems few years agom but it never
happened, and one reason (of many) is Novell policy with new releases.
-- -- Original Message -- --
From: "Vladimir Barac" <vbarac@(protected)>
To: <suse-oracle@(protected)>; <Arun.Singh@(protected)>
Sent: Monday, August 13, 2007 9:55 PM
Subject: [suse-oracle] More info on SLES10 needed
Hello to everyone
I have been following this thread about SLES10 problems, about RedHat
being preferred by dba's/sysadmins...
Could somebody give more details on problems with Suse SLES 10? We have
SLES10 deployed on number of servers, hosting 10.2 database - both rac
and single instance. So far we haven't seen any problems. However, after
reading some posts I became concerned so I will appreciate if somebody
could shed some light.
Thank you in advance,
Vladimir
__ ____ ____ ____ ____ ____ ____ ____ ____ ____ ____ ____ ____ ____ __
This email has been scanned by the MessageLabs Email Security System.
For more information please visit http://www.messagelabs.com/email
__ ____ ____ ____ ____ ____ ____ ____ ____ ____ ____ ____ ____ ____ __
--
To unsubscribe, email: suse-oracle-unsubscribe@(protected)
For additional commands, email: suse-oracle-help@(protected)
Please see http://www.suse.com/oracle/ before posting
|
|
