snipped to fit – apparently your messages to me are getting turned into huge pictures – I have no idea why. _____ From: Mark W. Farnham [mailto:mwf@xxxxxxxx] Sent: Tuesday, November 10, 2009 8:59 AM To: 'Nizar Ahmed'; 'hrishys@xxxxxxxxxxx'; 'oracle-l@xxxxxxxxxxxxx' Subject: RE: Backup VLDB's It absolutely is not a fact, nor is it written as a fact. It is a logical condition. Whether it has a logical value of TRUE is something you measure with each candidate backup technology against your actual load on your actual system. If a given backup process degrades your throughput or response time materially then you have to decide whether that degradation is tolerable, constrain the degradation to a window where it is not material, or reject that backup strategy. While tendancies of various backup methods to generate noticeable loads can be described, what works for a particular system has to be measured. Depending on whether journalling is used and other factors, for example, some varieties of split duplex or triplex backup have near zero overhead during the backup, but then put a significant load on the i/o subsytems until the re-imaging is complete. If someone measures only the overhead during the backup, then that is missed. Some varieties of splitting are quite efficient at both splitting and re-association. You have to measure. Your supposition that RMAN puts a noticeable load on overall system performance could be true for some systems and false for other systems. From the way you have written your response I think you have not measured yet. Your other point is very compelling, and is exactly what I am writing about when I say to design your recovery. If you already have a source from which to reload any lost data, then that may indeed represent your best recovery protocol if the reloading process can serve your operational requirements. Regards, mwf _____ From: Nizar Ahmed [mailto:gnahmed.c@xxxxxxxxxx] Sent: Tuesday, November 10, 2009 6:55 AM To: Mark W. Farnham; hrishys@xxxxxxxxxxx; oracle-l@xxxxxxxxxxxxx Cc: Nizar Ahmed Subject: RE: Backup VLDB's Thanks Hrishy and Mark for your comments. Fortunately I am at a place where $$ do not matter. EMC is our backup partner; we have VTL's; and are also trying out BCV to another node and backing up (RMAN) from the second node. This is for another system and this costs real money. We had EMC experts on site but this approach is still not stable / smooth as we would like it to be. Second though Mark's points have compelling logic – in the 5 years I have handled this system, I have not had to do a single recovery. And I am not stupid to extrapolate this to mean – "I don’t need a backup ever…" This system sources its data from various other systems and I can always fall back on them to reload lost data. See Mark's comments I have highlighted – Is this a guaranteed fact? Third – My question – "what's the fastest way to backup" is because I need read response time in seconds on these huge tables. I do not want the RMAN backup window to run for most of the day thereby putting a load on the overall system performance. We do use block change tracking and run a differential incremental on weekdays to keep the sizes of the backup small. Thanks gents,<snip> The information in this email may contain confidential material and it is intended solely for the addresses. Access to this email by anyone else is unauthorized. If you are not the intended recipient, please delete the email and destroy any copies of it, any disclosure, copying, distribution is prohibited and may be considered unlawful. Contents of this email and any attachments may be altered, Statement and opinions expressed in this email are those of the sender, and do not necessarily reflect those of Saudi Telecommunications Company (STC). قد يحتوي هذا البريد الالكتروني على معلومات سرية موجهة إلى الأشخاص المرسلة لهم فقط ولا يصرح لأي شخص آخر الاطلاع عليها، وفي حال استلام هذا البريد الالكتروني بشكل خاطئ فإنه يجب حذفه وإبلاغ المرسل بشكل مباشر. وأي تسريب لتلك المعلومات أو نسخها أو نشرها يعد أمراً مخالفاً وقد يؤدي إلى المسائلة القانونية، كما أن الآراء المذكورة بهذا البريد تمثل رأي مرسلها ولا تعبر بالضرورة عن رأي شركة الاتصالات السعودية