Sun cluster vs Veritas cluster
List, Just wondering if anyone on the list who had experience on both Sun clustering and Veritas clustering software share their experience on both products, strength and weakness or any comparison?? KC
Re: disk corruption?
Jerry, We had this problem recently on Solaris 2.6 and Oracle 8.0.5, Oracle initially said upgrade to the 8.0.5.2 patch will fix the problem, we ended up upgraded to the 8.0.6.2 to get rid of the problem. Anita is right about the datafile is still intact, it just a read error, you can run dbv to verify it. KC -Original Message- To: Multiple recipients of list ORACLE-L [EMAIL PROTECTED] Date: Friday, June 29, 2001 8:30 AM Jerry, If you're on 8.0.x and file 9 is 2GB then it's a known bug (774252) when a read crosses the 2GB boundary. The datafile is not corrupt, Oracle is merely unable to read it properly. It wasn't fixed in any of the patchsets, but there are PSE's (patchset exceptions) for most 8.0.x versions. Log a tar with support to get the right one for your specific RDBMS version. HTH, -- Anita --- Jerry C [EMAIL PROTECTED] wrote: Hi all, I'm getting a disk error: ORA-01115: IO error reading block from file 9 (block # 262144) ORA-01110: data file 9: '/orafiles/data01/xlepmprd/ewlarge.dbf' ORA-27072: skgfdisp: I/O error SVR4 Error: 25: Inappropriate ioctl for device Additional information: 262143 My sysadmin tells me that the disk is ok, since there are no errors in /var/adm/messages (I didn't see any, either). Is there a way to verify that the disk is free of errors?? Any other log files to look at? The platform is Sun: SunOS msubxm01 5.6 Generic_105181-12 sun4u sparc SUNW,Ultra-4 Thanks! Jerry _ Do You Yahoo!? Get your free @yahoo.com address at http://mail.yahoo.com -- Please see the official ORACLE-L FAQ: http://www.orafaq.com -- Author: Jerry C INET: [EMAIL PROTECTED] Fat City Network Services-- (858) 538-5051 FAX: (858) 538-5051 San Diego, California-- Public Internet access / Mailing Lists To REMOVE yourself from this mailing list, send an E-Mail message to: [EMAIL PROTECTED] (note EXACT spelling of 'ListGuru') and in the message BODY, include a line containing: UNSUB ORACLE-L (or the name of mailing list you want to be removed from). You may also send the HELP command for other information (like subscribing). __ Do You Yahoo!? Get personalized email addresses from Yahoo! Mail http://personal.mail.yahoo.com/ -- Please see the official ORACLE-L FAQ: http://www.orafaq.com -- Author: A. Bardeen INET: [EMAIL PROTECTED] Fat City Network Services-- (858) 538-5051 FAX: (858) 538-5051 San Diego, California-- Public Internet access / Mailing Lists To REMOVE yourself from this mailing list, send an E-Mail message to: [EMAIL PROTECTED] (note EXACT spelling of 'ListGuru') and in the message BODY, include a line containing: UNSUB ORACLE-L (or the name of mailing list you want to be removed from). You may also send the HELP command for other information (like subscribing). -- Please see the official ORACLE-L FAQ: http://www.orafaq.com -- Author: KC INET: [EMAIL PROTECTED] Fat City Network Services-- (858) 538-5051 FAX: (858) 538-5051 San Diego, California-- Public Internet access / Mailing Lists To REMOVE yourself from this mailing list, send an E-Mail message to: [EMAIL PROTECTED] (note EXACT spelling of 'ListGuru') and in the message BODY, include a line containing: UNSUB ORACLE-L (or the name of mailing list you want to be removed from). You may also send the HELP command for other information (like subscribing).
direct I/O with VxFS
Dear list, Our shop is running Solaris 2.6, Oracle database 8.0.6 and Veritas VxFS file system. At present, all database file access thru the file system buffer cache, we decided to use the direct I/O mount option with VxFS for performance reason. There is one issue I am not sure about and hope someone on the list can give me a pointer, access database file thu the file system buffer cache gives us read ahead advantage, with direct I/O I think what you read is what you get (no read ahead functionality), is this an issue?? If it is an issue, how do you deal with it?? WIth file system buffer cache, I think the database block size should be the same as file system buffer size and file system block size is irrelevant to how much system read or write, is this still true under direct I/O, I am not sure what dictate how much data to read or write?? KC
Re: direct I/O with VxFS
Connor, Thanks for your input. I agreed with you on what you said about the cache, however what I was asking is the read-ahead that the file system provided, for a physical read request, the file system actually read more blocks into the buffer, so the next physical read request can be satisfied from memory. I was wondering with direct I/O, is this still true?? Does VxFS read-ahead when it process a read request?? May be there is no such thing as read-ahead or it is irrelevant in this context, please correct me if my question doesn't make any sense! KC -Original Message- To: Multiple recipients of list ORACLE-L [EMAIL PROTECTED] Date: Thursday, June 28, 2001 1:27 AM The argument for direct io is that if you already have a large cache (the oracle one), then its not much use having a second copy of that cache (the unix one) - that memory could possibly be better used elsewhere (supporting more users, large sort sizes etc etc etc) So having direct io is giving more of the caching responsibility to the oracle buffer cache. hth connor --- KC [EMAIL PROTECTED] wrote: Dear list, Our shop is running Solaris 2.6, Oracle database 8.0.6 and Veritas VxFS file system. At present, all database file access thru the file system buffer cache, we decided to use the direct I/O mount option with VxFS for performance reason. There is one issue I am not sure about and hope someone on the list can give me a pointer, access database file thu the file system buffer cache gives us read ahead advantage, with direct I/O I think what you read is what you get (no read ahead functionality), is this an issue?? If it is an issue, how do you deal with it?? WIth file system buffer cache, I think the database block size should be the same as file system buffer size and file system block size is irrelevant to how much system read or write, is this still true under direct I/O, I am not sure what dictate how much data to read or write?? KC = Connor McDonald http://www.oracledba.co.uk (mirrored at http://www.oradba.freeserve.co.uk) Some days you're the pigeon, some days you're the statue Do You Yahoo!? Get your free @yahoo.co.uk address at http://mail.yahoo.co.uk or your free @yahoo.ie address at http://mail.yahoo.ie -- Please see the official ORACLE-L FAQ: http://www.orafaq.com -- Author: =?iso-8859-1?q?Connor=20McDonald?= INET: [EMAIL PROTECTED] Fat City Network Services-- (858) 538-5051 FAX: (858) 538-5051 San Diego, California-- Public Internet access / Mailing Lists To REMOVE yourself from this mailing list, send an E-Mail message to: [EMAIL PROTECTED] (note EXACT spelling of 'ListGuru') and in the message BODY, include a line containing: UNSUB ORACLE-L (or the name of mailing list you want to be removed from). You may also send the HELP command for other information (like subscribing). -- Please see the official ORACLE-L FAQ: http://www.orafaq.com -- Author: KC INET: [EMAIL PROTECTED] Fat City Network Services-- (858) 538-5051 FAX: (858) 538-5051 San Diego, California-- Public Internet access / Mailing Lists To REMOVE yourself from this mailing list, send an E-Mail message to: [EMAIL PROTECTED] (note EXACT spelling of 'ListGuru') and in the message BODY, include a line containing: UNSUB ORACLE-L (or the name of mailing list you want to be removed from). You may also send the HELP command for other information (like subscribing).
Re: What the difference between qio and direct I/O option in VxFS
Connor, You have confirmed what I believed to be true, you don't need the Veritas database edition for direct I/O. Just need to convince our SysAdmin now. KC -Original Message- To: Multiple recipients of list ORACLE-L [EMAIL PROTECTED] Date: Wednesday, June 20, 2001 5:44 PM Direct I/O is enabled simply be specifying in the mount options in /etc/vfstab (on Solaris). You can do it even with standard ufs volumes. As long as you have vxfs then you can also have vxfs with direct IO Quick I/O is the separately licensed bit I think - it should also outperform Direct I/O. Still - I'm a fan of raw over either... Cheers Connor --- KC [EMAIL PROTECTED] wrote: Connor, Thanks for the info. Is Direct I/O licensed with the Veritas Volume manager or the Database edition?? We want to be able to avoid double buffering, sound like Direct I/O is what we want, however our SysAdmin insisted that we need to purchase the Database edition, I thought Direct I/O came with the VxFS file system, maybe the SysAdmin confused Direct I/O to raw ,please clarify?? KC -Original Message- To: Multiple recipients of list ORACLE-L [EMAIL PROTECTED] Date: Wednesday, June 20, 2001 2:53 AM Direct IO is the capacity to perform operations without the data being passed through the Unix buffer cache (it goes straight to/from the disks to the Oracle buffer cache) Quick IO is Veritas's simulation of raw datafiles - it presents the file to Oracle as if it were raw. hth connor --- KC [EMAIL PROTECTED] wrote: Dear List, What the difference between qio and direct I/O option in Veritas file system VxFS?? KC = Connor McDonald http://www.oracledba.co.uk (mirrored at http://www.oradba.freeserve.co.uk) Some days you're the pigeon, some days you're the statue Do You Yahoo!? Get your free @yahoo.co.uk address at http://mail.yahoo.co.uk or your free @yahoo.ie address at http://mail.yahoo.ie -- Please see the official ORACLE-L FAQ: http://www.orafaq.com -- Author: =?iso-8859-1?q?Connor=20McDonald?= INET: [EMAIL PROTECTED] Fat City Network Services-- (858) 538-5051 FAX: (858) 538-5051 San Diego, California-- Public Internet access / Mailing Lists To REMOVE yourself from this mailing list, send an E-Mail message to: [EMAIL PROTECTED] (note EXACT spelling of 'ListGuru') and in the message BODY, include a line containing: UNSUB ORACLE-L (or the name of mailing list you want to be removed from). You may also send the HELP command for other information (like subscribing). -- Please see the official ORACLE-L FAQ: http://www.orafaq.com -- Author: KC INET: [EMAIL PROTECTED] Fat City Network Services-- (858) 538-5051 FAX: (858) 538-5051 San Diego, California-- Public Internet access / Mailing Lists To REMOVE yourself from this mailing list, send an E-Mail message to: [EMAIL PROTECTED] (note EXACT spelling of 'ListGuru') and in the message BODY, include a line containing: UNSUB ORACLE-L (or the name of mailing list you want to be removed from). You may also send the HELP command for other information (like subscribing). = Connor McDonald http://www.oracledba.co.uk (mirrored at http://www.oradba.freeserve.co.uk) Some days you're the pigeon, some days you're the statue Do You Yahoo!? Get your free @yahoo.co.uk address at http://mail.yahoo.co.uk or your free @yahoo.ie address at http://mail.yahoo.ie -- Please see the official ORACLE-L FAQ: http://www.orafaq.com -- Author: =?iso-8859-1?q?Connor=20McDonald?= INET: [EMAIL PROTECTED] Fat City Network Services-- (858) 538-5051 FAX: (858) 538-5051 San Diego, California-- Public Internet access / Mailing Lists To REMOVE yourself from this mailing list, send an E-Mail message to: [EMAIL PROTECTED] (note EXACT spelling of 'ListGuru') and in the message BODY, include a line containing: UNSUB ORACLE-L (or the name of mailing list you want to be removed from). You may also send the HELP command for other information (like subscribing). -- Please see the official ORACLE-L FAQ: http://www.orafaq.com -- Author: KC INET: [EMAIL PROTECTED] Fat City Network Services-- (858) 538-5051 FAX: (858) 538-5051 San Diego, California-- Public Internet access / Mailing Lists To REMOVE yourself from this mailing list, send an E-Mail message to: [EMAIL PROTECTED] (note EXACT spelling of 'ListGuru') and in the message BODY, include a line containing: UNSUB ORACLE-L (or the name of mailing list you want to be removed from). You may also send
Re: Disk configuration
Connor, It would be easy if I am the SysAdmin, our Mr. Unix here doesn't like raw just because he can't see the files in the file system. I am just exploring different ways to stripe the disk, I noticed that we made one subdisk for each disk and stripe the volume on them. This is quite different to the places I worked before, people create subdisks on the outer and inner tracks of the disk and stripe one volume on the outer tracks subdisk and another for the inner tracks, I assumed people do that for performance reason, any ideal on issue you will have for having one subdisk per disk? I agreed that raw is the best in term of performance. KC -Original Message- To: Multiple recipients of list ORACLE-L [EMAIL PROTECTED] Date: Wednesday, June 20, 2001 6:04 PM If you are going to that level for performance reasons, I would seriously consider using raw partitions to avoid the issue. hth connor --- KC [EMAIL PROTECTED] wrote: Kevin, Thanks for your input. I was trying to put certain datafiles on contiguous disk space, tell me if I am wrong, I try to avoid the situation where you want to create a 2G file, but the file system don't have a 2G contiguous space, so your flle is broken into multiple pieces, can that happen?? KC -Original Message- From: Kevin Lange [EMAIL PROTECTED] To: Multiple recipients of list ORACLE-L [EMAIL PROTECTED] Date: Wednesday, June 20, 2001 1:22 AM Subject: RE: Disk configuration It all depends on what kind of os/filesystem/and disks you have. I know that under AIX, using SSA drives we could actually tell where on the disk we wanted the filesystem to go. This way we could position certain things in the faster location. But personally, I would not go thru the trouble. I have never had a DB slowdown so far because of placement on the drive. Admittadly, I have had probelms based on putting conflicting tables/indexes on the same drive you want to keep things that could be access simultaneously on different media. But other than that no other conflicts. -Original Message- From: KC [mailto:[EMAIL PROTECTED]] Sent: Tuesday, June 19, 2001 9:36 AM To: Multiple recipients of list ORACLE-L Subject: Disk configuration Dear List, Someone told me when a disk receive a write request, it write to the nearest free space on disk where the disk read/write head is currently positioning, is this information correct?? If this is true, is this a bad thing for database application?? That mean we can't really control where the file go, for performance purpose we may want to put certain files on the outer tracks of a disk, if the write location is depending on where the read/write head is, how can we avoid that, can we create subdisks from the outer track of a disk and create a logical volume from it?? KC = Connor McDonald http://www.oracledba.co.uk (mirrored at http://www.oradba.freeserve.co.uk) Some days you're the pigeon, some days you're the statue Do You Yahoo!? Get your free @yahoo.co.uk address at http://mail.yahoo.co.uk or your free @yahoo.ie address at http://mail.yahoo.ie -- Please see the official ORACLE-L FAQ: http://www.orafaq.com -- Author: =?iso-8859-1?q?Connor=20McDonald?= INET: [EMAIL PROTECTED] Fat City Network Services-- (858) 538-5051 FAX: (858) 538-5051 San Diego, California-- Public Internet access / Mailing Lists To REMOVE yourself from this mailing list, send an E-Mail message to: [EMAIL PROTECTED] (note EXACT spelling of 'ListGuru') and in the message BODY, include a line containing: UNSUB ORACLE-L (or the name of mailing list you want to be removed from). You may also send the HELP command for other information (like subscribing). -- Please see the official ORACLE-L FAQ: http://www.orafaq.com -- Author: KC INET: [EMAIL PROTECTED] Fat City Network Services-- (858) 538-5051 FAX: (858) 538-5051 San Diego, California-- Public Internet access / Mailing Lists To REMOVE yourself from this mailing list, send an E-Mail message to: [EMAIL PROTECTED] (note EXACT spelling of 'ListGuru') and in the message BODY, include a line containing: UNSUB ORACLE-L (or the name of mailing list you want to be removed from). You may also send the HELP command for other information (like subscribing).
What the difference between qio and direct I/O option in VxFS
Dear List, What the difference between qio and direct I/O option in Veritas file system VxFS?? KC
Disk configuration
Dear List, Someone told me when a disk receive a write request, it write to the nearest free space on disk where the disk read/write head is currently positioning, is this information correct?? If this is true, is this a bad thing for database application?? That mean we can't really control where the file go, for performance purpose we may want to put certain files on the outer tracks of a disk, if the write location is depending on where the read/write head is, how can we avoid that, can we create subdisks from the outer track of a disk and create a logical volume from it?? KC
Re: What the difference between qio and direct I/O option in VxFS
Connor, Thanks for the info. Is Direct I/O licensed with the Veritas Volume manager or the Database edition?? We want to be able to avoid double buffering, sound like Direct I/O is what we want, however our SysAdmin insisted that we need to purchase the Database edition, I thought Direct I/O came with the VxFS file system, maybe the SysAdmin confused Direct I/O to raw ,please clarify?? KC -Original Message- To: Multiple recipients of list ORACLE-L [EMAIL PROTECTED] Date: Wednesday, June 20, 2001 2:53 AM Direct IO is the capacity to perform operations without the data being passed through the Unix buffer cache (it goes straight to/from the disks to the Oracle buffer cache) Quick IO is Veritas's simulation of raw datafiles - it presents the file to Oracle as if it were raw. hth connor --- KC [EMAIL PROTECTED] wrote: Dear List, What the difference between qio and direct I/O option in Veritas file system VxFS?? KC = Connor McDonald http://www.oracledba.co.uk (mirrored at http://www.oradba.freeserve.co.uk) Some days you're the pigeon, some days you're the statue Do You Yahoo!? Get your free @yahoo.co.uk address at http://mail.yahoo.co.uk or your free @yahoo.ie address at http://mail.yahoo.ie -- Please see the official ORACLE-L FAQ: http://www.orafaq.com -- Author: =?iso-8859-1?q?Connor=20McDonald?= INET: [EMAIL PROTECTED] Fat City Network Services-- (858) 538-5051 FAX: (858) 538-5051 San Diego, California-- Public Internet access / Mailing Lists To REMOVE yourself from this mailing list, send an E-Mail message to: [EMAIL PROTECTED] (note EXACT spelling of 'ListGuru') and in the message BODY, include a line containing: UNSUB ORACLE-L (or the name of mailing list you want to be removed from). You may also send the HELP command for other information (like subscribing). -- Please see the official ORACLE-L FAQ: http://www.orafaq.com -- Author: KC INET: [EMAIL PROTECTED] Fat City Network Services-- (858) 538-5051 FAX: (858) 538-5051 San Diego, California-- Public Internet access / Mailing Lists To REMOVE yourself from this mailing list, send an E-Mail message to: [EMAIL PROTECTED] (note EXACT spelling of 'ListGuru') and in the message BODY, include a line containing: UNSUB ORACLE-L (or the name of mailing list you want to be removed from). You may also send the HELP command for other information (like subscribing).
Re: Disk configuration
Chris, I agreed with what you said, however sometimes you want to put the file in certain location when the file is first created, any ideal?? KC -Original Message-From: Christopher Spence [EMAIL PROTECTED]To: Multiple recipients of list ORACLE-L [EMAIL PROTECTED]Date: Wednesday, June 20, 2001 1:15 AMSubject: RE: Disk configuration As that person if Santa exists. The datafiles allocate their extents upon their creation, so a new insert will write within that space, Walking on water and developing software from a specification are easy if both are frozen. Christopher R. Spence Oracle DBA Fuelspot -Original Message-From: KC [mailto:[EMAIL PROTECTED]]Sent: Tuesday, June 19, 2001 10:36 AMTo: Multiple recipients of list ORACLE-LSubject: Disk configuration Dear List, Someone told me when a disk receive a write request, it write to the nearest free space on disk where the disk read/write head is currently positioning, is this information correct?? If this is true, is this a bad thing for database application?? That mean we can't really control where the file go, for performance purpose we may want to put certain files on the outer tracks of a disk, if the write location is depending on where the read/write head is, how can we avoid that, can we create subdisks from the outer track of a disk and create a logical volume from it?? KC
Re: Disk configuration
Kevin, Thanks for your input. I was trying to put certain datafiles on contiguous disk space, tell me if I am wrong, I try to avoid the situation where you want to create a 2G file, but the file system don't have a 2G contiguous space, so your flle is broken into multiple pieces, can that happen?? KC -Original Message-From: Kevin Lange [EMAIL PROTECTED]To: Multiple recipients of list ORACLE-L [EMAIL PROTECTED]Date: Wednesday, June 20, 2001 1:22 AMSubject: RE: Disk configuration It all depends on what kind of os/filesystem/and disks you have. I know that under AIX, using SSA drives we could actually tell where on the disk we wanted the filesystem to go. This way we could position certain things in the faster location. But personally, I would not go thru the trouble. I have never had a DB slowdown so far because of placement on the drive. Admittadly, I have had probelms based on putting conflicting tables/indexes on the same drive you want to keep things that could be access simultaneously on different media. But other than that no other conflicts. -Original Message-From: KC [mailto:[EMAIL PROTECTED]]Sent: Tuesday, June 19, 2001 9:36 AMTo: Multiple recipients of list ORACLE-LSubject: Disk configuration Dear List, Someone told me when a disk receive a write request, it write to the nearest free space on disk where the disk read/write head is currently positioning, is this information correct?? If this is true, is this a bad thing for database application?? That mean we can't really control where the file go, for performance purpose we may want to put certain files on the outer tracks of a disk, if the write location is depending on where the read/write head is, how can we avoid that, can we create subdisks from the outer track of a disk and create a logical volume from it?? KC
Filesystem block size
Dear list, We are using Veritas filesystem, I try to find the filesystem block size by using fstype -v device and df -g, they both gave me 8K. Is the two commands reporting on the same thing or it is just happen to be the same?? KC
F45webm memory usage
Dear list, We are Oracle Apps site, one of the problem we are having is memory leak with f45webm processes. We had 4G of memory on the web server tier where the Form server run, when we run the prtmem command, it reported that 3.5G was used by user (application), using the memps command I can see the private area (I think this is the heap area of the process) of the f45webm is chewing up a lot of memory. We had opened tar with Oracle Support and applied all the patches suggested and also upgrade to the latest Form patchset, nothing seem to bring down the heap size. I further investigated and found something which I couldn't make sense of, I observed the paging rate and scan rate using sar throughout the day, there was not much happening, just a tiny bit of paging activities, if the box is using 3.5G of 4G, I would expected to see a lot of paging and scanning activities. Is anyone out there have similar problem with f45webm memory usage. I remembered reading something about Oracle server process do not release heap memory when it was done with it, it was up to the OS to reclaim it. In order words, the heap side shown is not really what is being used, sort of like a high water mark. Would the f45webm processes exhibits similar behaviour?? Also the prtmem and memps command, the number they displayed represent virtual or physical memory? I got a feeling is virtual because the total size included shared as well, but my SA said it is physical memory, any ideal?? KC
Difference in speed of sorting using index and non-index column in the order by
Dear list, Someone ask me when Oracle is doing a sort, will it make a difference if we use indexed column and non-indexes column in the order by clause, my initial thinking was index only speed up the retrieval of data from disk, Oracle probably using other algorithm in sorting records, so indexing is not relevant in that case. Can someone shed light on this question. KC
Re: TUSC and Kevin Loney
Just wondering if Steve joined TUSC, will there be no Steve Adams on the list and ixora on the net?? -Original Message- To: Multiple recipients of list ORACLE-L [EMAIL PROTECTED] Date: Wednesday, June 06, 2001 12:52 AM Hi All, No, I'm waiting for Henry to announce the details of the commercial DBA union! The pre-announcement is at http://www.lazydba.com/xcomment.pl?discussions:6 @ Regards, @ Steve Adams @ http://www.ixora.com.au/ @ http://www.christianity.net.au/ PS. This is a joke for the benefit of those who have been following the LazyDBA issues. -Original Message- Sent: Wednesday, 6 June 2001 0:35 To: Multiple recipients of list ORACLE-L What about Steve Adams, does somebody know whether he joined some big company also? -- Please see the official ORACLE-L FAQ: http://www.orafaq.com -- Author: Steve Adams INET: [EMAIL PROTECTED] Fat City Network Services-- (858) 538-5051 FAX: (858) 538-5051 San Diego, California-- Public Internet access / Mailing Lists To REMOVE yourself from this mailing list, send an E-Mail message to: [EMAIL PROTECTED] (note EXACT spelling of 'ListGuru') and in the message BODY, include a line containing: UNSUB ORACLE-L (or the name of mailing list you want to be removed from). You may also send the HELP command for other information (like subscribing). -- Please see the official ORACLE-L FAQ: http://www.orafaq.com -- Author: KC INET: [EMAIL PROTECTED] Fat City Network Services-- (858) 538-5051 FAX: (858) 538-5051 San Diego, California-- Public Internet access / Mailing Lists To REMOVE yourself from this mailing list, send an E-Mail message to: [EMAIL PROTECTED] (note EXACT spelling of 'ListGuru') and in the message BODY, include a line containing: UNSUB ORACLE-L (or the name of mailing list you want to be removed from). You may also send the HELP command for other information (like subscribing).
async I/O with DB writer and LOG writer
Dear list, I recently discussed the effect of using async I/O with DB writer and LOG writer with someone, I was told that async I/O is more beneficial to DB writer than LOG writer, what is your opinion?? Can the difference be quantify?? I thought aysnc I/O with DB writer can be compensated by multiple DB writers, is there any comparison between the two?? KC
OFF TOPIC: load average figure in TOP
Dear list, The load average figure shown in top, does it represent average number of running jobs or running jobs + jobs ready to be run. One of my machine had 4 CPU, sometimes the load average is 10, so I guess the load average number is not just running jobs. Any ideal?? Kam
OFF TOPIC: How to find lib search order for executable under Sun solaris
Dear list, In HP-UX, one can do a chatr to find out the lib search order of an executable, is there a corresponding command under Sun Solaris. Also, can someone tell me how nmliblist is used when relink Oracle under Sun?? Thanks Kam
Unix: File cache size on Solaris
Dear List, We are running Oracle on Solaris 2.6, I am new to Solaris env and wondering anyone can explain to me on the following: We have 6G of memory on the Sun E5500 server, when I do a prtmem command, I notice the size for File Cache is about 3.9G big and 100M free space. I was curious about what is taking up all the space, our Unix admin explained to me that under Solaris, the OS will grep any free memory as unix buffer and give it back to process when needed and we can't configure how much memory Solaris keep as unix buffer. I remember under HP-UX, you can keep a limit on how much memory the OS can keep under buffer cache. In our case, will such a large pool of buffer cache has any performance impact, Solaris must spend a lot of time maintaining this pool and also it probably very expensive to allocate and deallocate buffer between OS and user processes. Any ideal?? KC
