From a Linux OS, is there a way to determine if the OS is running as a guest OS on a virtualized VMware environment as opposed to running directly on non-virtualized/bare metal/physical host? In my case it is either VMware or not, but I am also interested the more general question of whether the OS is on physical hardware or virtualized hardware of any sort.
We have a folder on a windows server shared to provide access to an important application. When the application is being updated we (DBAs) change the permissions on the share to deny all access to it and then disconnect any files open on the share. After we complete the application update (using a different share to the same files) we re-enable access to the share.
The way we modify the share is using the MS Management Console to remotely connect to the server. Apparently this capability requires administrator access on the server or at least some level of permissions that the Infrastructure team would like to take away.
The question is how can the DBAs handling the application update disable and enable access to the share with as few permissions on the server as possible?
On our Windows 2008 R2 64 bit servers there seems to be a discrepancy between the number shown in Task Manager on the Performance tab in the Memory Graph and the sum of the Memory (Private Working Set) on the Processes tab. The former is consistently higher. The discrepancy does not seem to depend on the amount of memory allocated to the box or on the number of processes on the box. For the four servers I have examined the discrepancy ranged from 853 MB to 4,178 MB. Can someone please explain this?
We have a VSphere server and increased the drive for the C:\ drive on one of the VMs from 20GB to 30GB. Disk management shows the correct new size in the Disk view, but the volume information still shows 20GB as does the Local Disk (C:) Properties. We have re-scanned and refreshed several times and even rebooted the VM. Anyone have any ideas?
We have several relatively recent (less than 1 week) installations of Windows Server 2008 R2 Enterprise that are getting this error whenever we open Disk Management. The application name is vdsldr.exe and the Problem Event Name is BEX64.
The windows firewall was turned off and the service disabled. The Virtual Disk service was started (identical results when it is stopped). The computers are virtual machines running on VSphere, so that could be part of the problem.
The strange thing is that Disk Manager worked fine when the VMs were first created.
Task Manager also seems to be crashing on these machines. The problems are occuring on multiple VMs that are on different physical hosts.
We use Oracle's ORDCOM infrastructure on our 11.1.0.7 database to interact with a COM object. The COM object doesn't do anything with the database. This process has been working fine for several years.
Today I created an Oracle Wallet and added credentials for sys to so that a script I have can login using "/@coredev as sysdba" rather than store the username and password in clear text.
The problem is that as soon as I added the line "SQLNET.WALLET_OVERRIDE = TRUE" to the sqlnet.ora file, the ORDCOM OACreate call raises an ORA-28578: protocol error during callback from an external procedure. I have a program outside of the database that can test the COM object and that continues to work correctly. As soon as I comment out the line in sqlnet.ora it all works correctly, but of course I can no longer use the Wallet.
I have an Oracle 11.1.0.7 database running on windows 32 bit. I would like to upgrade to 64 bit (same version of oracle) and would like to use a standby database for a rolling upgrade. Is this possible and can I make the standby the primary so I don't have to have another outage to roll back to the primary?
We are using Automatic Memory Management on Oracle 11.1.0.7. Last week Friday the Shared Pool unexpectedly increased from ~600MB to ~850MB. The memory came from the buffer cache which went from ~400 to ~150. Needless to say the system is doing significantly more disk reads than before. So far I haven't been able to prove that this was an incorrect change, but it sure doesn't seem right to me. I'd like to know what caused it so I can hopefully reverse it.
I did an AWR comparison from before and after and I noticed that the Shared Pool increase was almost exclusively in the sql area. I looked at the SQL being stored there using queries such as the following, but didn't find anything that would have changed recently:
--Replace anything in quotes and numbers with % so that statements not using
--bind variables group together.
SELECT count(*), Trunc(sum(sharable_mem)/1024/1024) SharableMemory,
regexp_replace(regexp_replace(
sql_text,'.\''.+\''','''%'''),'[0-9]+','%') || ''';' sql
FROM v$sql
GROUP BY regexp_replace(regexp_replace(
sql_text,'.\''.+\''','''%'''),'[0-9]+','%')
HAVING sum(sharable_mem) > 2*1024*1024
ORDER BY Trunc(sum(sharable_mem)/1024/1024) DESC;
I also tried flushing the shared pool. The system doesn't give up any of the shared pool to the buffer cache, perhaps because the size increases too rapidly after the flush.
I also tried setting a minimum size for the buffer cache to force it to give up the memory from the shared pool. On my test system this method allows me to increase the buffer cache by about 100MB, but on my live system it won't even increase 3MB before giving me an error indicating there isn't enough memory to perform the operation.
The alert log shows "Sweep Incident[77273]: completed" near the time when the change happened. I couldn't find much information on metalink about it. It also shows that DIA0 was restarted near that time, but I don't see how that could be related.
Here are several potentially unrelated notes. The AWR showed almost double the number of pin requests for all Namespaces in the Library Cache than there were before the problem started. The database was upgraded from 10.2.0.4 about a month ago. The number of sessions increased from ~200 to ~250 about a week before the problem occurred.
I'm leaving an instance restart as a last resort since I don't know if it will help.
Any suggestions would be appreciated.
We have some automated software that checks certain Exchange 2007 mailboxes via POP3. When a customer's incoming e-mail requests a read receipt, Exchange seems to be sending back a non-read receipt.
Not read: <subject>
Your message was deleted without being read on...
The order in which the software issues POP3 commands is
USER
PASS
STAT
UIDL
TOP # 0
LIST #
RETR #
TOP # 0
RSET
DELE #
QUIT
Our exchange administrator has been able to duplicate this issue via raw POP3 commands over telnet, port 110. The commands he ran were simply
USER
PASS
STAT
RETR #
DELE #
QUIT
Does Exchange 2007 translate POP3 commands to read flags? If so, it should tell that I've issued a RETR command to retrieve the message. Is this a bug? Is there a way to disable the sending of read receipts for specific mailboxes or globally?
For our 10g Oracle database I had a script that would periodically create batch files to run DBVerify on all our datafiles. This way if I needed to check for corruption I would just have to run the batch files. After upgrading to Oracle 11g my batch files do not worked. I traced the problem down to dbv itself. Here is the command generated by the script:
dbv FILE=F:\ORACLE\ORADATA\CD\CAMPUSCHURCH1.DBF BLOCKSIZE=4096
LOGFILE='c:\DBVerify_COREDEV\CAMPUSCHURCH1_55.txt'
When I run this it gives a DBV-00100: Specified FILE (C:\Windows\system32/F:\ORACLE\ORADATA\CD\CAMPUSCHURCH1.DBF) not accessible. Note: The folder I am in is C:\Windows\system32. I can get it to work by switching to F:\ and removing F:\ from my dbv call, but I don't like that because then I'd have to change my script to switch to the proper drive before each dbv call.
Can anyone verify my findings and/or provide a solution?
The Windows Services control panel by default opens showing the console tree on the left and in extended mode which consumes additional space on the left. In order to view as much information as possible the first thing I do is click the toolbar button to hide the console and click the Standard tab at the bottom and expand the name column. I know I could do this with a macro, but I was wonder if anyone knew of a way to do any of these things automatically, perhaps with a registry change or a command line parameter.
Windows Server 2008 apparently allows an application to somehow configure the folder so that any changes made within the folder require administrator level access. I login with an account that has administrator privileges, but is not the local administrator account. When I do so I find that I can't save changes to files opened within this folder. I know I can open the application as administrator or move the file out of the folder, make the change, then move it back in, but I'm hoping there is a better way short of disabling the protection entirely. Is there a way perhaps to remove it for the files I frequently edit?
For an Intel based system keeping the number of cores constant how much difference would there be between two dual core processors and one quad core processor? We run Oracle OLTP databases on two dual core processors. With fewer dual core processors available we may need to switch to quad core, but we are limited to four cores due to licensing issues, so will have to switch to one processor. If we make the switch should I expect to see a change in real world processing speed?