It's a very common scenario: some process wants to drop a file on a server every 30 minutes or so. Simple, right? Well, I can think of a bunch of ways this could go wrong.
For instance, processing a file may take more or less than 30 minutes, so it's possible for a new file to arrive before I'm done with the previous one. I don't want the source system to overwrite a file that I'm still processing.
On the other hand, the files are large, so it takes a few minutes to finish uploading them. I don't want to start processing a partial file. The files are just tranferred with FTP or sftp (my preference), so OS-level locking isn't an option.
Finally, I do need to keep the files around for a while, in case I need to manually inspect one of them (for debugging) or reprocess one.
I've seen a lot of ad-hoc approaches to shuffling upload files around, swapping filenames, using datestamps, touching "indicator" files to assist in synchronization, and so on. What I haven't seen yet is a comprehensive "algorithm" for processing files that addresses concurrency, consistency, and completeness.
So, I'd like to tap into the wisdom of crowds here. Has anyone seen a really bulletproof way to juggle batch data files so they're never processed too early, never overwritten before done, and safely kept after processing?
The key is to do the initial juggling at the sending end. All the sender needs to do is:
Store the file with a unique filename.
As soon as the file has been sent, move it to a subdirectory called e.g. completed.
Assuming there is only a single receiver process, all the receiver needs to do is:
Periodically scan the completed directory for any files.
As soon as a file appears in completed, move it to a subdirectory called e.g. processed, and start working on it from there.
Optionally delete it when finished.
On any sane filesystem, file moves are atomic provided they occur within the same filesystem/volume. So there are no race conditions.
Multiple Receivers
If processing could take longer than the period between files being delivered, you'll build up a backlog unless you have multiple receiver processes. So, how to handle the multiple-receiver case?
Simple: Each receiver process operates exactly as before. The key is that we attempt to move a file to processed before working on it: that, and the fact the same-filesystem file moves are atomic, means that even if multiple receivers see the same file in completed and try to move it, only one will succeed. All you need to do is make sure you check the return value of rename(), or whatever OS call you use to perform the move, and only proceed with processing if it succeeded. If the move failed, some other receiver got there first, so just go back and scan the completed directory again.
If the OS supports it, use file system hooks to intercept open and close file operations. Something like Dazuko. Other operating systems may let you know about file operations in anoter way, for example Novell Open Enterprise Server lets you define epochs, and read list of files modified during an epoch.
Just realized that in Linux, you can use inotify subsystem, or the utilities from inotify-tools package
File transfers is one of the classics of system integration. I'd recommend you to get the Enterprise Integration Patterns book to build your own answer to these questions -- to some extent, the answer depends on the technologies and platforms you are using for endpoint implementation and for file transfer. It's a quite comprehensive collection of workable patterns, and fairly well written.
Related
I have to read the MFT file of a running Windows (XP or higher) and through it to reach the HD sectors that held the contents ($DATA) of a specific file that exists on the machine.
The problem is that between the time of reading the MFT until the fetching of the relevant sectors and reading them, the file system structure can vary and the locations may not be relevant anymore.
Is there a way to "freeze" the system for a certain time? Perhaps guarantee that there will not be changes for this file? Lock a specific file in order to make it not moving between sectors? (Including due to optimizations and changes in indirect)
Of course I would prefer not to copy the entire hard disk and to work statically since it's a slow operation that would disallow normal use of the system at this time. Needless to say, I don't want to use the API functions of the OS or to write a driver.
I'd simply open the file, requesting read/write access, with read share mode. If you succeed to open the file, you're guaranteed that data will not change until you close the handle. See https://msdn.microsoft.com/en-us/library/windows/desktop/hh449422%28v=vs.85%29.aspx
If you want to achieve that on files that are already opened and locked by different processes, that's entirely different story and I believe you have to write own filter driver.
If the file location in the system varies, it will be accordingly reflected in the MFT. So instead of trying to stop any activity for the file you can simply compare the MFT info before and after reading the file. Unless you are de-fragmenting or deleting contents of the file the file storage structure will not change. Additions to files do not affect the consistency of data that you read. So if this is your scenario, you can just go ahead with the above method.
My app is monitoring a "hot" folder somewhere on the local filesystem for newly added files to push to a network location. I'm running into a problem when very large files are being written into the hot folder: the file system event notifying me of changes in the hot folder will fire well before the file completes writing. When my app tries to upload the file, it mis-reads the file size as the current number of copied bytes, not the eventual total number of bytes.
Things I've tried:
NSURL getResourceValue:forKey:error: to read NSURLAllocatedFileSizeKey (same value as NSURLFileSizeKey while the file is being written).
NSFileManager attributesOfItemAtPath:error: to look at NSFileBusy (always NO).
I can't seem to find any mechanism short of repeatedly polling a file for its size to determine if the file is finished copying and can be uploaded.
There aren't great ways to do this.
If you can be certain that the writer is using NSFileCoordinator, then you can also use that to coordinate your access to the file.
Likewise, if you're sure that the writer has opted in to advisory locking, you could try to open the file for shared access by calling open() with the O_SHLOCK and O_NONBLOCK flags. If you succeed, then there are no other descriptors open for exclusive access. You can either use the file descriptor you've got or close it and then use some other API to access the file.
However, if you can't be sure of any of those, then your best bet may be to set a timer to repeatedly check the file's metadata (size, date modified, etc.). Only when you see that it has stopped changing over a reasonable time interval (2 seconds, maybe) would you attempt to access it (and cancel the timer).
You might want to do all three. Wait for the file's metadata to settle down, then use a NSFileCoordinator to read from the file. When it calls your reader block, use open() with O_SHLOCK | O_NONBLOCK to make sure there are no other processes which have exclusive access to it.
You need some form of coordinated file locking.
fcntl() and flock() are common functions for this.
Read up on it first.
Then see what options you have.
If you can control the code base of those other processes, all the better.
The problem with really large files is that what's changed or changing inside them is opaque and isn't always at the end.
Good processes should generally be doing atomic writes. (Write to a temp file then swap it out) but if these files are actually databases then you will want to look at using the db's server app for this sort of thing.
If the files are wrappers containing other files then it gets extra messy as those contents might have dependencies on one another to be in a usable state.
Every once in awhile I am fed a large data file that my client uploads and that needs to be processed through CMFL. The problem is that if I put the processing on a CF page, then it runs into a timeout issue after 120 seconds. I was able to move the processing code to a CFC where it seems to not have the timeout issue. However, sometime during the processing, it causes ColdFusion to crash and has to restarted. There are a number of database queries (5 or more, mixture of updates and selects) required for each line (8,000+) of the file I go through as well as other logic provided by me in the form of CFML.
My question is what would be the best way to go through this file. One caveat, I am not able to move the file to the database server and process it entirely with the DB. However, would it be more efficient to pass each line to a stored procedure that took care of everything? It would still be a lot of calls to the database, but nothing compared to what I have now. Also, what would be the best way to provide feedback to the user about how much of the file has been processed?
Edit:
I'm running CF 6.1
I just did a similar thing and use CF often for data parsing.
1) Maintain a file upload table (Parent table). For every file you upload you should be able to keep a list of each file and what status it is in (uploaded, processed, unprocessed)
2) Temp table to store all the rows of the data file. (child table) Import the entire data file into a temporary table. Attempting to do it all in memory will inevitably lead to some errors. Each row in this table will link to a file upload table entry above.
3) Maintain a processing status - For each row of the datafile you bring in, set a "process/unprocessed" tag. This way if it breaks, you can start from where you left off. As you run through each line, set it to be "processed".
4) Transaction - use cftransaction if possible to commit all of it at once, or at least one line at a time (with your 5 queries). That way if something goes boom, you don't have one row of data that is half computed/processed/updated/tested.
5) Once you're done processing, set the file name entry in the table in step 1 to be "processed"
By using the approach above, if something fails, you can set it to start where it left off, or at least have a clearer path of where to start investigating, or worst case clean up in your data. You will have a clear way of displaying to the user the status of the current upload processing, where it's at, and where it left off if there was an error.
If you have any questions, let me know.
Other thoughts:
You can increase timeouts, give the VM more memory, put it in 64 bit but all of those will only increase the capacity of your system so much. It's a good idea to do these per call and do it in conjunction with the above.
Java has some neat file processing libraries that are available as CFCS. if you run into a lot of issues with speed, you can use one of those to read it into a variable and then into the database
If you are playing with XML, do not use coldfusion's xml parsing. It works well for smaller files and has fits when things get bigger. There are several cfc's written out there (check riaforge, etc) that wrap some excellent java libraries for parsing xml data. You can then create a cfquery manually if need be with this data.
It's hard to tell without more info, but from what you have said I shoot out three ideas.
The first thing, is with so many database operations, it's possible that you are generating too much debugging. Make sure that under Debug Output settings in the administrator that the following settings are turned off.
Enable Robust Exception Information
Enable AJAX Debug Log Window
Request Debugging Output
The second thing I would do is look at those DB queries and make sure they are optimized. Make sure selects are happening with indicies, etc.
The third thing I would suspect is that the file hanging out in memory is probably suboptimal.
I would try looping through the file using file looping:
<cfloop file="#VARIABLES.filePath#" index="VARIABLES.line">
<!--- Code to go here --->
</cfloop>
Have you tried an event gateway? I believe those threads are not subject to the same timeout settings as page request threads.
SQL Server Integration Services (SSIS) is the recommended tool for complex ETL (Extract, Transform, and Load) work, which is what this sounds like. (It can be configured to access files on other servers.) The question might be, can you work up an interface between Cold Fusion and SSIS?
If you can upgrade to cf8 and take advantage of cfloop file="" which would give you greater speed and the file would not be put in memory (which is probably the cause of the crashing).
Depending on the situation you are encountering you could also use cfthread to speed up processing.
Currently, an event gateway is the only way to get around the timeout limits of an HTTP request cycle. CF does not have a way to process CF pages offline, that is, there is no command-line invocation (one of my biggest gripes about CF - very little offling processing).
Your best bet is to use an Event Gateway or rewrite your parsing logic in straight Java.
I had to do the same thing, Ben Nadel has written a bunch of great articles uses java file io, to allow you to more speedily read files, write files etc...
Really helped improve the performance of our csv importing application.
I'm running a very computationally intensive scientific job that spits out results every now and then. The job is basically to just simulate the same thing a whole bunch of times, so it's divided among several computers, which use different OSes. I'd like to direct the output from all these instances to the same file, since all the computers can see the same filesystem via NFS/Samba. Here are the constraints:
Must allow safe concurrent appends. Must block if some other instance on another computer is currently appending to the file.
Performance does not count. I/O for each instance is only a few bytes per minute.
Simplicity does count. The whole point of this (besides pure curiosity) is so I can stop having every instance write to a different file and manually merging these files together.
Must not depend on the details of the filesystem. Must work with an unknown filesystem on an NFS or Samba mount.
The language I'm using is D, in case that matters. I've looked, there's nothing in the standard lib that seems to do this. Both D-specific and general, language-agnostic answers are fully acceptable and appreciated.
Over NFS you face some problems with client side caching and stale data. I have written an OS independent lock module to work over NFS before. The simple idea of creating a [datafile].lock file does not work well over NFS. The basic idea to work around it is to create a lock file [datafile].lock which if present means file is NOT locked and a process that wants to acquire a lock renames the file to a different name like [datafile].lock.[hostname].[pid]. The rename is an atomic enough operation that works well enough over NFS to guarantee exclusivity of the lock. The rest is basically a bunch of fail safe, loops, error checking and lock retrieval in case the process dies before releasing the lock and renaming the lock file back to [datafile].lock
The classic solution is to use a lock file, or more accurately a lock directory. On all common OSs creating a directory is an atomic operation so the routine is:
try to create a lock directory with a fixed name in a fixed location
if the create failed, wait a second or so and try again - repeat until success
write your data to the real data file
delete the lock directory
This has been used by applications such as CVS for many years across many platforms. The only problem occurs in the rare cases when your app crashes while writing and before removing the lock.
Why not just build a simple server which sits between the file and the other computers?
Then if you ever wanted to change the data format, you would only have to modify the server, and not all of the clients.
In my opinion building a server would be much easier than trying to use a Network file system.
Lock File with a twist
Like other answers have mentioned, the easiest method is to create a lock file in the same directory as the datafile.
Since you want to be able to access the same file over multiple PC the best solution I can think of is to just include the identifier of the machine currently writing to the data file.
So the sequence for writing to the data file would be:
Check if there is a lock file present
If there is a lock file, see if I'm the one owning it by checking that its content has my identifier.
If that's the case, just write to the data file then delete the lock file.
If that's not the case, just wait a second or a small random length of time and try the whole cycle again.
If there is no lock file, create one with my identifier and try the whole cycle again to avoid race condition (re-check that the lock file is really mine).
Along with the identifier, I would record a timestamp in the lock file and check whether it's older than a given timeout value.
If the timestamp is too old, then assume that the lock file is stale and just delete it as it would mea one of the PC writing to the data file may have crashed or its connection may have been lost.
Another solution
If you are in control the format of the data file, could be to reserve a structure at the beginning of the file to record whether it is locked or not.
If you just reserve a byte for this purpose, you could assume, for instance, that 00 would mean the data file isn't locked, and that other values would represent the identifier of the machine currently writing to it.
Issues with NFS
OK, I'm adding a few things because Jiri Klouda correctly pointed out that NFS uses client-side caching that will result in the actual lock file being in an undetermined state.
A few ways to solve this issue:
mount the NFS directory with the noac or sync options. This is easy but doesn't completely guarantee data consistency between client and server though so there may still be issues although in your case it may be OK.
Open the lock file or data file using the O_DIRECT, the O_SYNC or O_DSYNC attributes. This is supposed to disable caching altogether.
This will lower performance but will ensure consistency.
You may be able to use flock() to lock the data file but its implementation is spotty and you will need to check if your particular OS actually uses the NFS locking service. It may do nothing at all otherwise.
If the data file is locked, then another client opening it for writing will fail.
Oh yeah, and it doesn't seem to work on SMB shares, so it's probably best to just forget about it.
Don't use NFS and just use Samba instead: there is a good article on the subject and why NFS is probably not the best answer to your usage scenario.
You will also find in this article various methods for locking files.
Jiri's solution is also a good one.
Basically, if you want to keep things simple, don't use NFS for frequently-updated files that are shared amongst multiple machines.
Something different
Use a small database server to save your data into and bypass the NFS/SMB locking issues altogether or keep your current multiple data files system and just write a small utility to concatenate the results.
It may still be the safest and simplest solution to your problem.
I don't know D, but I thing using a mutex file to do the jobe might work. Here's some pseudo-code you might find useful:
do {
// Try to create a new file to use as mutex.
// If it's already created, it will throw some kind of error.
mutex = create_file_for_writing('lock_file');
} while (mutex == null);
// Open your log file and write results
log_file = open_file_for_reading('the_log_file');
write(log_file, data);
close_file(log_file);
close_file(mutex);
// Free mutex and allow other processes to create the same file.
delete_file(mutex);
So, all processes will try to create the mutex file but only the one who wins will be able to continue. Once you write your output, close and delete the mutex so other processes can do the same.
Me and my colleague are trying to implement a mechanism to get recovery from broken files on an embedded equipment.
This could be happened during certain circumstances, e.g. user takes off the battery during file writing.
Orz, but now we have just one idea:
Create duplicated backup files, and copy them back if dangerous file i/o is not finished properly.
This is kind of stupid, as if the backup files also broken, we are just dead.
Do you have any suggestions or good articles on this?
Thanks in advance.
Read up on database logging and database journal files.
A database (like Oracle) has very, very robust file writing. Do not actually use Oracle. Use their design pattern. The design pattern goes something like this. You can borrow these ideas without actually using the actual product.
Your transaction (i.e., Insert) will fetch the block to be updated. Usually this is in memory cache, if not, it is read from disk to memory cache.
A "before image" (or rollback segment) copy is made of the block you're about to write.
You change the cache copy, write a journal entry, and queue up a DB write.
You commit the change, which makes the cache change visible to other transactions.
At some point, the DB writer will finalize the DB file change.
The journal is a simple circular queue file -- the records are just a history of changes with little structure to them. It can be replicated on multiple devices.
The DB files are more complex structures. They have a "transaction number" -- a simple sequential count of overall transactions. This is encoded in the block (two different ways) as well as written to the control file.
A good DBA assures that the control file is replicated across devices.
When Oracle starts up, it checks the control file(s) to find which one is likely to be correct. Others may be corrupted. Oracle checks the DB files to see which match the control file. It checks the journal to see if transactions need to be applied to get the files up to the correct transaction number.
Of course, if it crashes while writing all of the journal copies, that transaction will be lost -- not much can be done about that. However, if it crashes after the journal entry is written, it will probably recover cleanly with no problems.
If you lose media, and recover a backup, there's a chance that the journal file can be applied to the recovered backup file and bring it up to date. Otherwise, old journal files have to be replayed to get it up to date.
Depends on which OS etc. etc. but in most cases what you can do is copy to a temporary file name and as the last final step rename the files to the correct name.
This means the (WOOPS) Window of Opertunity Of Potential S****p is confined to the interval when the renames take place.
If the OS supports a nice directory structure and you lay out the files intelligently you can further refine this by copying the new files to a temp directory and renaming the directory so the WOOPS becomes the interval between "rename target to save" and "rename temp to target".
This gets even better if the OS supports Soft link directories then you can "ln -s target temp". On most OSes replacing a softlink will be an "atomic" operation which will work or not work without any messy halfway states.
All these options depend on having enough storage to keep a complete old and new copy on the file system.