Does anyone know a way to configure a Grid Control notification rule for a failure of the SQL APPLY process on a logical standby database? I'd really like an email notification that the SQL APPLY process has encountered an error and is no longer applying redo to the logical standby.
Help, please!
Hardware: Dell Poweredge 2600
PERC 4/Di
4 SCSI Drives, 1 standalone
3 in a RAID 5 configuration
OS: Windows 2000 Server
In other words, a fairly old system. Anyway, we are in the process of taking over support for this site. The current tech wants out and is fading from view fast, so we need to solve this problem:
The standalone disk (where the OS was) failed. We've replaced the disk, installed the OS, but need to know exactly how to proceed from here. I've never worked with a RAID system before, so I don't want to touch anything without knowing what I'm doing.
We are not certain if the site will want us to attempt to recover the array or wait for the old tech to become available. We have replaced the server with a temporary box, and recovered MOST of the data from an online backup service. However, the other tech failed to backup a part of the data and the only copy of it is on this RAID array. Hence, our caution.
We have poked minimally around in the boot-up PERC config utility, and it seems to me that that's where we'll need to be to reclaim the array. Another possibility is that there is some Dell software for the RAID controller we need to acquire.
Can anyone provide clues as to how to proceed from here? Any help GREATLY appreciated.
Oracle 10gR2, Win2003.
I know enough about RMAN to perform basic functions. I'm trying to learn more.
We have an RMAN backup process that is used to duplicate a database nightly. The backup is generated on a single disk. The script allocates 2 channels for that backup, both pointing to the same directory. It works fine, generating 2 backupsets (backupsets? I think. I'm a bit confused about terminology: a backup consists of one or more backupsets?). Anyway, the duplicate process picks up these backupsets and restores a copy of the database from them. I noticed today that it allocates 3 channels, but only two appear to be used. Is that because the backup process only used 2? Which brings me to the major point: given my configuration, is there any advantage to allocating more channels for backup and restore? This is an 80GB database, and the restores takes over 2 hours to complete which seems slow to me - I can do a hot backup in that amount of time, and I thought RMAN had performance advantages over simple scripted backups.
To summarize my questions:
- Does an RMAN backup consist of one or more backupsets?
- Can a restore operation only use as many channels as were used in the backup?
- Is there any advantage to increasing the number of channels for both given a backup to a single disk?
10gR2 Database, in Data Guard configuration to both a physical and logical standby.
I have a materialized view that does a complete refresh periodically of about 40K records from a remote database. It's been working on an hourly refresh for years without issue. Yesterday I was asked to make the refresh rate considerably more frequent, like, every 2 minutes. I did that, no apparent problems, until this morning around 4am the archive log directory filled up.
I've been struggling all day with trying to turn off logging. I have:
- set NO FORCE LOGGING in the database
- altered the materialized view and base table to NOLOGGING
- altered the indexes on the view to NOLOGGING
Nothing seems to reduce the log generation rate. The view is refreshed in a scheduled task via dbms_refresh.refresh right before the procedure that reads it is kicked off.
I know I should revamp the process, probably by sucking the data into a global temp table each time the job runs, but that's going to take a while to get through the QA mill. Any insight as to what I am missing here? From what I can read it seems I should be able to do this. This asktom article seems to support that, but I can't seem to make it work.
Thanks so much for your time.
I've got a Windows 7 Ultimate system which maps a Samba 3.0.33 share. I have no problems reading, writing, or replacing any files or directories from windows file explorer. I have attempted to set up a push/pull sync script on the windows client using xcopy. I am trying to replace any files newer than the corresponding file/directory in either direction. The copy from the server to the client works great. The copy from the client to the server fails with "access denied" (when the directory already exists - new directories can be created without issue).
This is my windows script:
xcopy c:\source_dir z:\dest_dir /D /E /I /F /R /Y
xcopy z:\dest_dir c:\source_dir /D /E /I /F /R /Y
Here's what I see for the first command:
C:\Source_dir>xcopy c:\Source_dir\test z:\Dest_dir\test /E /I /F /R /Y
Access denied
Unable to create directory - Z:\Dest_dir\test
0 File(s) copied
The same command works fine on an XP system connected to the same server. There must be something I'm missing - any ideas what it might be?
Thanks!
PS: I forgot to mention that if the directory does not exist already on the server, the xcopy command to the server succeeds.
PPS: Robocopy yields identical results.
I've had Windows 7 Ultimate 32 bit installed on my desktop for a week or so now, and there's one pretty big issue for me. My home network has a Centos 5.3 file server running Samba 3.0.33. Right next to it, hooked to the same switch, I've got a W2K server. So I get file transfer speeds from the W2K box to the W7 box around 1.5MB/s, while access to the Centos Box only gets 250KB/s. It's not just via Samba, either - ftp exhibits the same problem. I've tried a new NIC in the W7 box. I've tried putting the W7 hostname in the /etc/hosts file on the Centos box. I've swapped cables from the switch to the servers. None of these had any noticeable effect. Any ideas on what else to check? If I were paranoid I'd think Microsoft was sabotaging connectivity with Linux ;-) One more thing: before installing W7, I had an XP Pro install that worked perfectly fine. I really don't think it's hardware at this point.
Oracle 10.2.0.4 database with a logical standby on Win2K3. Recently a rather large delete operation was carried out on the production instance. I'm experiencing difficulty with the logical standby, in that it gets a couple of hundred (58M size) archive logs into the operation and the apply process fails with an out-of-memory error. Unfortunately, every time it fails it has to restart the apply from the beginning of the transaction. This is taking a couple of days each time. Anyway, in trying to resolve this problem, I've noticed that each archive log from the production system generates 5 or 6 log switches on the standby. I don't understand why this should be. Anyone have any ideas?
A related question that I've not found the answer for: does anyone know if the logical standby must be running in archivelog mode? I really don't have a need to keep the logs.
I have an Oracle logical standby database being managed via data guard. Just this morning the redo apply process began failing with an ORA-01919 error, indicating one of our application roles did not exist. However, I can see the role on both primary and standby databases. We also have a physical standby that has long since applied the redo where this is happening on the logical, without issue.
I have opened an SR with Oracle. I was wondering if anyone out there has seen this before.
I guess I should mention: Oracle 10.2.0.4, Win2003 Server SP2.
UPDATE: So far, Oracle Support has not provided an answer. I thought I'd post here what I have learned so far.
It appears that a grant of DBA on the primary host to a role works fine for users granted the role. It does not work on the logical standby. IOW:
create role TEST;
grant dba to TEST;
grant TEST to auser;
connect auser
set role TEST;
grant <existing role> to <existing user>;
This works on the primary instance but fails on the logical. A workaround appears to be to grant each role on the primary to the role TEST with admin option in the logical:
grant <existing role> to TEST with admin option; <== do this on the logical standby
Then the command works on the logical standby.
So I know I've had this working before, but after a fresh install of Windows XP and Cygwin, I can't seem to get the Cygwin/X server to give me anything other than a gray screen and an hourglass cursor. I followed the instructions on the Cygwin/X website. I haven't been able to find anything helpful via Google. Anyone here seen this?
I left work Friday with a long running SQL Tuning Advisor session running from Grid Control in Oracle, on one of our development instances. Some of these sessions take over 4 hours to run, so I prefer to run them on clones of production rather than the production box. My problem is I would like to implement the profiles at times on the production instance. Is there a way to do that?