File Lock with NFS cache

I use the API FileChannel.lock() to get the file lock for gobal lock based on NFS system, I know that many operations on the NFS system are not atomic because of system cache, but I also know the NFS system has its own lock mechanism, so I don't know if the API FileChannel.lock() is safe as the global lock, which accessed by processes from different nodes.
OS Linux 2.6.16.60-0.21-smp
NFS server v3
NFS Client v3

Crossposted and answered.

Similar Messages

File locking with OSX Server & Microsoft Word

We have a small office LAN based on a Airport Extreme bases station. Periodically Microsoft Word reports a break in connection to the shared folder where we store documents to get disconnected. When that happens and the user reconnects to the share, the user gets the file read-only message for the Word document that was open at the time.
I have found that in order to clear the read-only flag, I have to restart the server and open the Word file on the server with the ID of the user who was editing the file on the remote computer. Only that seems to clear the file locking that causes the file to be read-only.
I cannot figure out a less drastic way to release the lock file. I don't see any hidden temp files in the same folder as the document that is locked for editing found by ls -a or in the root direction of the shared folder in the .TemporaryItems/folders.<UID> that seems to be holding the lock.
I have run chflags nouchg <Word file> from Terminal, but that seems to have no effect either. I also used xattr to see if there were any locks held that way, but I don't see anything that way either.
File sharing to Macs only, which are configured just to use AFP through OSX Server.
Does anyone understand how the file locks of Word documents can be released directly with out restarting the server?
I should add that the underlying problems seems to be WiFi related. The behavior is that the WiFi connection seems to break long enough that the shared volumes disconnects. Outlook also causes a break. This behavior seems to have started with 10.9 and the purchase of new Retina Display MacBook Pros. I turned off AppNap on the Office applications but that is not clearly helpful. I also have been told the problem seems more likely to occur when Word is open in the background and another program like Outlook is in the foreground.
Any suggestions appreciated.

Apple write an operating system and also produce file sharing software as part of that to be used as a file server. Apple provide documentation for third-party software developers on how they should write software to work with Apple's software and also give those developers early access to new versions so those third-party developers can test and if needed make adjustments and issue updates to cope with any Apple changes.
Some third-party developers are good at dealing with this, some are bad, and some totally ignore what Apple does and give the impression they don't care if their product works properly or not. I think we can all judge where Microsoft sits.
It appears Microsoft have never paid any attention in particular to how Apple expect file-locking to be handled when accessing files on a Mac server. There have been problems for years and years and years with Office. Two other problems I have seen which seem different to yours but probably related are -
With Office documents it is supposed to be possible for more than one person to edit the same file at the same time, consider it a miracle if this actually works
With Office in particular Word, there is an auto-save function, unfortunately the way this seems to be implemented it seems that Word creates a new temporary file each time it auto-saves the document and keeps all the previous temporary files open still, this eventually means potentially over a hundred temporary files are open - just for one Word document and you can then hit a limit on the total number of files you are allowed to have open at the same time. At this point further auto-saves fail, and you also encounter great difficulty doing a real full save of the document.
I do not hold Apple completely blameless over this issue, it is likely their file-locking implementations change too often, and have inadequate documentation, however even considering this a company the size of Microsoft with the amount of sales (and profit) they derive from Office for the Mac has no excuses at all for failing to put the effort in to resolve any such clearly critical problems.
We could go on and on about other areas where Microsoft don't play by the (software) rules. Even in Windows Office does not obey the standard print dialog rules Microsoft specified themselves!
Unfortunately not only is Office for Mac upgraded infrequently, but even when new paid for upgrades are released Microsoft have a history of still not fixing bugs. It goes without saying that a mere free update is even less likely to actually fix a bug, typically such free updates only address security issues. The next version of Office is going to be Office for Mac 2014 see http://www.macworld.com/article/2106643/microsoft-will-release-a-new-version-of- office-for-mac-this-year.html
One area I confidently predict Microsoft not to resolve in Office for Mac 2014 is that fact that Word for Mac still does not support right-to-left languages like Hebrew and Arabic. This is despite the fact that OS X itself has supported this for years and years, and despite the fact other Mac programs support this including the free TextEdit and Pages - both of which can read Word files. Some people may remember that at one point the Israeli Government temporarily banned all Microsoft software over this issue. See http://apple-beta.slashdot.org/story/03/10/15/2215249/israeli-government-suspend s-microsoft-contracts This issue goes back over TEN years!!
I note that Microsoft has now taken their OfficeForMac blog offline, probably due to the weight of criticisms. I would not say it is due to out-right anti-Microsoft hate, that war ended long ago. We just want products that work. I myself do use Microsoft products, even at home - where they are the best solution, so I use Microsoft Media Center for example. Sadly this is now being neglected by Microsoft.

Mandatory File Locking In Nfs V3 : How To?

I have a central file say, "central.txt ", mounted on an NFS v3 server. Now I have a minimum of 20 processes wanting to update the "central.txt" file at the same time. To ensure that the entries of this "central.txt" is consistent, I should lock the file before I update it and unlock the file after i update it.
I read is some where that NFS v3 supports only advisory locking. I need a mandatory locking system here. Is there any way I can issue a "mandatory lock" on the central.txt file?
I tried the following without any help.
**NOTE** : I'm using perl to write my script (that uses the central.txt file). It is this script that I run concurrently.
1. flock : doesn't work
2. system ("lockfile -l 10 -s 8 filename); : doesn't work
3. fcntl : doesn't work
4. File::NFSLock : doesn't work
Could any one tell me how to obtain a mandatory lock on a file mounted in NFS v3 and also how to trap any error while locking the file?
Thanks

Can you give more details on your environment? Anyway, you could put them in a database and have that take care of the locking matters if that's feasible in your situation.

Basic use of locks with replicated cache

Hi,
I'm in the process of evaluating Coherence and I have a few fundamental questions about best locking practices with a simple replicated cache. I've been through the docs & forums a few times, but I'm still confused.
The docs say that a replicted cache is "fully coherent". What, exactly does this mean? Does this imply that Coherence handles all the locking for you? Under what situations do I have to lock a node?
Thanks.

Hi,
There are generally two reasons for desiring full synchronicity:
Fault tolerance ... however, note that by the time the put returns, the data is held by both the client and the issuer. Also, we recommend the use of the Distributed cache topology for transactional data (and not Replicated). Distributed is in fact fully synchronous.
The other reasons people usually ask about this is concern over the logical view of data updates (simultaneous state across the cluster). Even within a single JVM running on a single CPU, data updates are not simultaneous or even synchronous. This is quite intentional (in fact most of the complexity of the Java Memory Model arises from the desire to avoid the requirement for simultaneous state changes).
In the JVM, synchronized blocks are required to avoid race conditions and inconsistent views of data, and in Coherence, locks are required to do the same.
The key point to remember is, from the point of view of a single cluster member, data updates are in fact synchronous and simultaneous (with or without locking).
I'm assuming your question relates to one of those two issues (as those are the most common concerns that we hear); if not, could you provide a bit more background on your requirements?
Jon Purdy
Tangosol, Inc.

File by File picking with Standard File Adapter with NFS Protocal !!

Hi Folks,
How to pick file by file with little gap in a file directory by Standard File Adapter. One interface, one communication channle.
Ex: XML_XXXXX_XXXX1.xml
      XML_XXXXX_XXXX2.xml
      XML_XXXXX_XXXX3.xml
First I need to pick one file XM_XXXXX_XXXX1.xml then wait for 2 minutes and pick another one in ascending order.
I have tried with Advance option "File size" and "Msec to wait before modification check" with 1200000 MS, but no use.
Please folks any one can help in this, which is very big help?
Kind Regards
Many Thanks in advance
San

Hi
My BPM steps are as follows:
1. Receive step -- Asynch
2. Transformation -- split the message
3. Switch step -- Condition "ID != ' ' from transform message'
4. In side Switch Branch 1 -- Receiver determination - Receive acknowledgment - Application
5. Send 1 -- with Application Transport, Queue Name - 'XXX'
      This XXX queue I have configured on ID with EOIO Service
6. block step - Send multiple messages to same receiver on top
    I have send 2 different message to one receiver, the first message is only one , second message are multiple. I am splitting    with transformation step
7. Sender -2 no acknowledgement active
8. stop
Can you please tell me where I am gonna wrong!!
I will apprecite for your efforts
Regards
San

File Locking Issue with Concurrent Locks

Hi. I'm trying to implement a procedure using the File Lock class for two instances of the same program (running on different servers) accessing files shared over an NFS mount in Unix (this is using Java 1.5). 99% of the time, this works great. One box will get the lock, and the other will respect that lock and not touch the file. However, the other 1% of the time, both boxes seem to access the same file simultaniously, allowing BOTH to acquire locks on the same file.
I'm wondering if anyone has any advice on how to solve this problem. I've been searching all over but haven't found a solution. Like I said, the locking works most of the time - I've even independantly verified that it does (and that it's not just timing) with a test program. It's only when both programs access the file at the exact same does that there seems to be a problem.
If it helps, here is the gist of my code:
      FileLock javaLock = null;
      FileChannel fileChannel = null;
      for(int i = 0; i < fileArray.length; i++)
         javaLock = null;
         fileChannel = null;
         if(!fileArray.exists())
continue;
try
RandomAccessFile raLockFile = new RandomAccessFile(fileArray[i], "rw");
try
fileChannel = raLockFile.getChannel();
if(fileArray[i].exists())
javaLock = fileChannel.tryLock();
catch(OverlappingFileLockException ofle)
// Error printed here
continue;
catch(Exception e)
// Error
javaLock = null;
if(javaLock == null)
continue; // File is already locked, move to next one
// Processing is done on the file here
catch(Exception e)
// Error
finally
try
if(javaLock != null)
javaLock.release();
if(fileChannel != null)
fileChannel.close();
catch(Exception e)
// Error
Edited by: kksmith on Jul 21, 2008 7:31 AM
Edited by: kksmith on Jul 21, 2008 7:32 AM

I don't have the errors on hand at the moment, but most were File Not Found IOExceptions (because one box would finish processing the file and rename it out from under the other box). Using a bunch of System.out.println()s I was able to determine that both boxes were locking the same specific files producing those errors.
The other weird thing that would happen is that the file would be processed correctly on one box, but somehow leave behind a zero length version of itself, which the other box would finish processing normally as if we had received a zero length file (as we have special code in place for that condition). This is probably the biggest problem, since I can trap File Not Founds (though it's messy and kind of defeats the purpose of using locks to begin with), but I don't have a real easy way to catch this (and I can't just toss out all zero lengths because we want to know if we legitamently got a zero length file in).
As far as I can tell though, this problem doesn't occur outside of an NFS mount, but I haven't tested it extensively enough to be sure (since it's a random timing thing, it's difficult to reproduce outside of just dumping a ton of files on it). I'm not entirely sure how our NFS system in configured, since that is completely out of my hands (System Admins have contol over all of that). But I believe it has most of the latest patches.

File name with symbols won't delete from trash.cache\trash\cache folder.

found this weird file name with symbols (squares nad the like) in the trash.cache\trash\cache folder. Can't seem to delete it from windows, can't get at from the dos prompt. Windows safe mode won't delete it.
Any suggestions as to what it is and how to get rid of it.
At present am trying reinstall of firefox and virus scan.
Thanks
Peter

I tried to do the instructions Adobe gave me but when
I put in the disc that came with my mac and hold down
C when it restarts it takes me to the screen to do a
fresh install.
At that point go to the Menu & select Disk Utility - I can't remember exactly which menu but you should be able to find it easily... there isn't too much there
I went into the disk utility through the Apps folder
and for somereason the option to repair isnt
highlighted and it wont let me click it. I tried to
repair permissions/verify but it doesnt change
anything. I looked at the info and it says the volume
can be repaired, but it wont let me.
You can't Repair the disk the system is currently running the OS from - That's why you have to boot from the Installer disk (or some other start-up disk). Repair Disk addresses directory structure issues - totally separate from what Repair Permissions does.
HTH|:>)
Bob J.

Office 2010 & 2007 - Excel and Access File Locking Out On the Network With Multiple Users

This is also posted in the Office 2010 - IT Pro General Discussions, but was suggested to repost here, since a definitive answer was not found.
Hi,
An issue that's happening is that Excel and Access files are locking on the network. We're currently using Office 2007 and 2010.
Here are some different scenarios that are happening:
When opening the file it is locked out by “User X” which is the person that has the file locked out and no one else can open the file.
When opening the file it is locked out by “User Y” which is NOT the actual person, but is locked out by “User X” and no one else can access the file.
When opening the file it is locked out by “…another user” which is generic and no one else can access the file.
The two more common events are incident 1 and 2 with 3 happening the less common.
This message will continue until the sessions are closed through computer management on the file server.
The file server is running Windows Server 2003.
This does happen on both Windows XP and Windows 7 clients.
This does happen for users using Office 2007 and Office 2010.
There are two sets of Office 2010 Users when it comes to patches. Everyone has the most current patches with Office 2010 SP2 while anyone that has Microsoft Project 2010 is using all the current update before Office 2010 SP2.
All users that are using Office 2007 have all the current patches and service packs.
Another variable is that we have users that will leave a file open on the network for 3+ days and after a while it will lock the file out.
Also we have Shadow Copy that runs daily on the system which I'm not for sure if that impacts anything if a file is opening during that time.
Any ideas on how to mitigate the lock out issues would be appreciated.
Thanks,
Binary Process
Edit November 12, 2013: This issue can occur if and if not another person actually has the file open. If the person doesn't have the file open then there is a hung connection which needs to be disconnected by going to the Computer Management of the File
server.

Hi Binary,
I know that the description of the hotfix does not relate to the issue. The purpose is to install it for upgrading SMB related file.
A similar issue I encountered before:
http://social.technet.microsoft.com/Forums/windowsserver/en-US/b7fcc59b-52d9-4a02-863a-1a529bcb8cb1/temp-doc-etc-files-dont-close-after-a-file-closes-this-causes-locked-files?forum=winserverfiles
It is resolved by upgrading SMB files so maybe it will help on your case.
Another hotfix which may related:
http://support.microsoft.com/kb/983458
If you have any feedback on our support, please send to [email protected]

File locked message when saving a file with Photoshop

We've been seeing a problem with several users that is hard to reproduce. Every once in a while when they try to save a file in Photoshop on the server they get a dialogue saying 'the file is locked, use Finder info to unlock the file'.
This file is located on the server and the problem only happens when doing a 'Save', not with a 'Save as'. The server is running 10.5.5 and the clients are 10.4.11. This is happening with Photoshop versions CS1, 2 and 3. It also happend with Illustrator CS3 once.
We've been experimenting with different scenario's using different ACL's and POSIX settings. Also, when the client reboots their machine the exact same 'Save' actions works again.

I've been receiving complaints about a similar issue over the past few weeks, except using Quark and Excel. I wonder if it's related to the 10.5.5 update. Are you sure you didn't get the version numbers reversed in your post? We're having the issue with 10.4.11 SERVER and 10.5.5. CLIENTS.
It's a file locking issue, so you probably won't fix it by changing the ACLs or POSIX permissions. The server thinks a different process still has the file open, so it's preventing you from changing the file. "Save As" works because it creates a new file.
You can probably eject & reconnect the volume from the client instead of fully rebooting.
The question is what's keeping the file open. You can run:
sudo lsof | grep <filename>
... in Terminal to see whether a given Mac has the file open. If the server comes back with a response to that command (don't worry what it says - you just want to see if anything comes back), then you know that the file is locked.
Next, go to each client that could have the file open and run the same command. If none of them have it open, you're probably looking at the same glitch that I've been. The server thinks a file is open but none of the clients do.

Problem with NFS file adapter using a network share as source of files

Hello,
I have set up a sender cc with NFS format file adapter, with at the source directory a network share (
server\sharename ). The share is on another system and not mapped to a local drivename, but it is open for everybody.
The problem is, the adapter seems to not pick up the files in the directory, so there must be something I'm overlooking.
- Both servers run on Windows
- Share is available normally and open for everybody for all actions (including the SAPServiceSID user)
- Files themselves are okay. If I put a file on a local drive and make the source directory the local drive instead of the network path, the file gets processed correctly
- Because of the need for a file completion check, FTP is not an option
Does anybody know what is going wrong, or what I could check for trace / logging info on possible cause of the problem?

Hi,
While using NFS mode you are supposed to put the file on XI server directory NOT to a common share directory of any other machine. And this is the reason your adapter in not picking the file.
You can upload the file on XI server directory by using the tcode "SXDA_TOOLS", morevere you can check the file on the same directory with tcode "AL11". It is otherway-around, you can first select or chose the dir by AL11 and then put the file in the same dir by tcode SXDA_TOOLS.
If you need more info about SXDA_TOOLS, let me know.
Regards,
Sarvesh

When I open my RAW files in Photoshop CC they are grainy and the JPEGS are not. What is going on? It doesn't seem to have anything to do with my CACHE levels either.

When I open my RAW files in Photoshop CC they are grainy and the JPEGS are not. What is going on? It doesn't seem to have anything to do with my CACHE levels either.

BOILERPLATE TEXT:
Note that this is boilerplate text.
If you give complete and detailed information about your setup and the issue at hand,
such as your platform (Mac or Win),
exact versions of your OS, of Photoshop (not just "CS6", but something like CS6v.13.0.6) and of Bridge,
your settings in Photoshop > Preference > Performance
the type of file you were working on,
machine specs, such as total installed RAM, scratch file HDs, total available HD space, video card specs, including total VRAM installed,
what troubleshooting steps you have taken so far,
what error message(s) you receive,
if having issues opening raw files also the exact camera make and model that generated them,
if you're having printing issues, indicate the exact make and model of your printer, paper size, image dimensions in pixels (so many pixels wide by so many pixels high). if going through a RIP, specify that too.
A screen shot of your settings or of the image could be very helpful too,
etc.,
someone may be able to help you (not necessarily this poster, who is not a Windows user).
Please read this FAQ for advice on how to ask your questions correctly for quicker and better answers:
http://forums.adobe.com/thread/419981?tstart=0
Thanks!

Solaris 10 NFS caching problems with custom NFS server

I'm facing a very strange problem with a pure java standalone application providing NFS server v2 service. This same application, targeted for JVM 1.4.2 is running on different environment (see below) without any problem.
On Solaris 10 we try any kind of mount parameters, system services up/down configuration, but cannot solve the problem.
We're in big trouble 'cause the app is a mandatory component for a product to be in production stage in a while.
Details follows
System description
Sunsparc U4 with SunOS 5.10, patch level: Generic_118833-33, 64bit
List of active NFS services
disabled   svc:/network/nfs/cbd:default
disabled   svc:/network/nfs/mapid:default
disabled   svc:/network/nfs/client:default
disabled   svc:/network/nfs/server:default
disabled   svc:/network/nfs/rquota:default
online       svc:/network/nfs/status:default
online       svc:/network/nfs/nlockmgr:default
NFS mount params (from /etc/vfstab)
localhost:/VDD_Server - /users/vdd/mnt nfs - vers=2,proto=tcp,timeo=600,wsize=8192,rsize=8192,port=1579,noxattr,soft,intr,noac
Anomaly description
The server side of NFS is provided by a java standalone application enabled only for NFS v2 and tested on different environments like: MS Windows 2000, 2003, XP, Linux RedHat 10 32bit, Linux Debian 2.6.x 64bit, SunOS 5.9. The java application is distributed with a test program (java standalone application) to validate main installation and configuration.
The test program simply reads a file from the NFS file-system exported by our main application (called VDD) and writes the same file with a different name on the VDD exported file-system. At end of test, the written file have different contents from the read one. Indeep investigation shows following behaviuor:
_ The read phase behaves correctly on both server (VDD) and client (test app) sides, trasporting the file with correct contents.
_ The write phase produces a zero filled file for 90% of resulting VDD file-system file but ending correctly with the same sequence of bytes as the original read file.
_ Detailed write phase behaviour:
1_ Test app wites first 512 bytes => VDD receive NFS command with offset 0, count 512 and correct bytes contents;
2_ Test app writes next 512 bytes => VDD receive NFS command with offset 0, count 1024 and WRONG bytes contents: the first 512 bytes are zero filled (previous write) and last 512 bytes with correct bytes contents (current write).
3_ Test app writes next 512 bytes => VDD receive NFS command with offset 0, count 1536 and WRONG bytes contents: the first 1024 bytes are zero filled (previous writes) and last 512 bytes with correct bytes contents (current write).
4_ and so on...
Further tests
We tested our VDD application on the same Solaris 10 system but with our test application on another (Linux) machine, contacting VDD via Linux NFS client, and we don�t see the wrong behaviour: our test program performed ok and written file has same contents as read one.
Has anyone faced a similar problem?
We are Sun ISV partner: do you think we have enough info to open a bug request to SDN?
Any suggestions?
Many thanks in advance,
Maurizio.

I finally got it working. I think my problem was that I was coping and pasting the /etc/pam.conf from Gary's guide into the pam.conf file.
There was unseen carriage returns mucking things up. So following a combination of the two docs worked. Starting with:
http://web.singnet.com.sg/~garyttt/Configuring%20Solaris%20Native%20LDAP%20Client%20for%20Fedora%20Directory%20Server.htm
Then following the steps at "Authentication Option #1: LDAP PAM configuration " from this doc:
http://docs.lucidinteractive.ca/index.php/Solaris_LDAP_client_with_OpenLDAP_server
for the pam.conf, got things working.
Note: ensure that your user has the shadowAccount value set in the objectClass

Sessions hangs with library cache lock

Der all,
11.1.0.7 rac on solaris 10
Our workflow session yesterday hanged on a particular step the session was waiting on a library cache lock (by using query select event,p1,p2 from v$session where sid=<my_sid>;)
when I checked the blocking session , using the note 122793.1 and http://oracle-study-notes.blogspot.com/2009/05/resolving-library-cache-lock-issue.html and http://oracle-study-notes.blogspot.com/2009/05/find-session-holding-library-cache-lock.html .
I found that
SQL> SELECT SID,USERNAME,TERMINAL,PROGRAM FROM V$SESSION
2   WHERE SADDR in
3    (SELECT KGLLKSES FROM X$KGLLK LOCK_A
4     WHERE KGLLKREQ > 0
5       AND EXISTS (SELECT LOCK_B.KGLLKHDL FROM X$KGLLK LOCK_B
                 WHERE KGLLKSES = '&SADDR_OF_BLKING_SESS'
6    7                   AND LOCK_A.KGLLKHDL = LOCK_B.KGLLKHDL
8                   AND KGLLKREQ = 0)
9    );
Enter value for saddr_of_blking_sess: 0000000770E494E0
old   6:                  WHERE KGLLKSES = '&SADDR_OF_BLKING_SESS'
new   6:                  WHERE KGLLKSES = '0000000770E494E0'
SID USERNAME        TERMINAL   PROGRAM
817 SYS             UNKNOWN    oracle@tabsdb07
                                  (J002)
828 SYS             UNKNOWN    oracle@tabsdb07
                                  (J001)after killing the session, the library cache locks still remained.when I ran trace on the session
select /*+ all_rows ordered */ A.rowid, :1, :2, :3
from
"DBMRPT"."DBM_BIAUTO_SUSP" A , "DBMRPT"."DBM_CDR_FILE_HEAD" B where(
"A"."CDR_TYPE" is not null and "A"."FILE_ID" is not null) and(
"B"."CDR_TYPE" (+)= "A"."CDR_TYPE" and "B"."FILE_ID" (+)= "A"."FILE_ID")
and( "B"."CDR_TYPE" is null or "B"."FILE_ID" is null)
call     count       cpu    elapsed       disk      query    current        rows
Parse        1      0.01       0.01          0          0          0           0
Execute      1      0.00       0.00          0          0          0           0
Fetch        0      0.00       0.00          0          0          0           0
total        2      0.01       0.01          0          0          0           0
Misses in library cache during parse: 1
Misses in library cache during execute: 1
Optimizer mode: ALL_ROWS
Parsing user id: SYS   (recursive depth: 3)
Elapsed times include waiting on following events:
Event waited on                             Times   Max. Wait Total Waited
----------------------------------------   Waited ---------- ------------
row cache lock                                  5        0.00          0.00
db file sequential read                   295932636        0.07       5066.63
gc cr grant 2-way                          727813        0.02        233.95
latch: gc element                              80        0.00          0.00
latch: gcs resource hash                      870        0.00          0.00
latch free                                      2        0.00          0.00
gc remaster                                     9        2.00         12.91
gcs drm freeze in enter server mode             9        0.54          2.08
latch: object queue header operation           66        0.00          0.05
latch: cache buffers chains                    15        0.03          0.20
resmgr:internal state change                   63        0.10          5.30
latch: cache buffers lru chain               1260        0.00          0.01
********************************************************************************Please guide
Kai
all this time sql_id for the session remanined in the sql :
ALTER TABLE DBMRPT.DBM_BIAUTO_SUSP ENABLE CONSTRAINT DBS1_DCFH_FK ..

hi..
Go through [http://orainternals.wordpress.com/2009/06/02/library-cache-lock-and-library-cache-pin-waits/]
Anand

File Locks and SMB shares with ML

I've been doing a lot of research on SMB and the way it locks files during access. I've made a lot of ground work in my research but could use a little further assistance from this support community.
The symptom is simple to explain: Users are occasionally being prompt for a username and password when attempting to rename or move files and folders.
After doing some research on this topic, I have discovered that there is a direct relationship to files being open on the server at the time the user is attempting to rename or move the folder. The following thread, albiet old, appears to have nailed the problem on the head http://arstechnica.com/civis/viewtopic.php?p=24558131. In particular, there appears to be file locking happening when preview is turned on through the finder. I've had all users remove preview from their Macs and this appears to have helped reduce the occurances of the password prompt, but has not completly solved the problem. This is also a work around, not a fix.
I've been using a series of command to help me trace the problem including openfile.exe on the Windows 2012 Storage Server (sharing the files via smb only) to discover who has what files open on the server, and the lsof command on the client workstations to discover what process has the file open. So far, the finder is consistently the only thing with the file open... even with the finder preview turned off. I've also found that the "open file" is simply the fact that the offending users Mac has a finder window with just the folder open (none of the files within the folder or previews open).
Is the real solution to simply close all finder windows when you're done working in a folder, or is there more that anyone can think of to help me find out exactly what is holding the file lock? Is this a known bug in the SMB implementation of ML? Can we expect to see a fix with Mavericks which will now be using SMB2.0?
Any help or information anyone can provide would be greatly appreciated. I have a bunch of documentation on this issue and would be happy to share. Please let me know if anyone needs any additional details.

Hello all - many apologies for my delay in posting here; and Squiggle, thanks for the second-hand nudge. As Stephen said we have been testing a solution concerning a setting in Finder's view options for the last few months, and these seem to have solved the issue in hand.
Essentially, we found that Finder was holding files open whenever the 'Show Icon Preview' option was set, on any of the four folder views, on any client machine accessing the share. Below is a piece of documentation I wrote up for our Service desk explaining how to diagnose and manually fix this issue on the client:
        - In Finder, open any folder
        - Click on the cog icon and select 'show view options.' Check that, in the dialogue box which appears, the 'Show icon preview' box is not checked. Click the other three Finder views and check it's turned off here too.
        - Click 'Use as Defaults'
In order to make this change remotely on multiple machines, you will need to change the clients' com.apple.plist files, and set every instance of the <showIconPreview> key to <False>. This is nested within several key / dictionary pairs in com.apple.finder.plist, once under <standardViewOptions> and thrice within <standardViewSettings>. As Stephen has already mentioned, the fact that this key is nestled deep within compound dictionaries seems to render them untouchable by defaults, though I would be very happy to be corrected on this.
How you push this change out will depend on your management system. We had been using Casper to
to create a managed preference pertaining to <standardViewOptions> and <standardViewSettings> within com.apple.finder.plist. These contained as values the entire dictionary associated with these keys, with the value for each <showIconPreview> set to false. This was then applied at a User Enforced Level (running every logon after Finder has set up the system defaults).
As a side note, I have found that certain machines (Such as the 10.8.3 machine I'm working on now) contain a key named <FK_StandardViewSettings>, which I have been unable to ascertain the purpose or origination of. These don't seem to affect the fix, so we've left them alone.
Irritatingly, Casper has dropped support for custom Managed Preferences in their latest release, so this problem has now resurfaced. I will keep this page updated with any fixes or workarounds I find.
Hope that helps,
Josh Smith

File Locking on solaris with C++ fstream

Hi Guys,
I have a problem with file locking on Solaris 5.6 with fstream. All of the file locking methods I know act on a file descriptor (flock,lockf, fcntl). In this particular scenario I got to use any C++ streams.
I'm particularly concerned with fstream filebuf flags. In Window NT Iam using filebuf::sh_none falg to prevent another process or thread from accessing the same file till I explicitly close the file.
I read some where, we can use filebuf::openprot on solaris. But If I use this openprot my Compiler(Work Shop 5.0) giving compilation error
Error: openprot is not a member of std::basic_filebuf<char, std::char_traits<<char>>.
So, if somebody can provide me with some other options about this it would be very helpful.
Any help greatly appreciated
Thanks

Sorry to be playing Twenty Questions, but I see
a problem with Solaris support. potentally. What
version of Solaris are you using? You might need
to use a later version of FD to avoid potential
bugs interacting with the OS.

File Lock with NFS cache

Similar Messages

Maybe you are looking for