The server has only 64MB of memory. I'm trying to push a huge git repository to it. Initially the target directory contains an empty bare repository. The push fails:
$ git push server:/tmp/repo master
Counting objects: 3064514, done.
Compressing objects: 100% (470245/470245), done.
fatal: Out of memory, calloc failed
error: pack-objects died of signal 13
error: failed to push some refs to 'server:/tmp/repo'
$ ssh server cat /tmp/repo.git/config
[pack]
threads = 1
deltaCacheSize = 8m
windowMemory = 32m
[core]
repositoryformatversion = 0
filemode = true
bare = true
I get the same error message after changing git config pack.windowMemory 16m on the server.
The same push succeeds to localhost:
$ git push 127.0.0.1:/tmp/repo master
Password:
Counting objects: 3064514, done.
Compressing objects: 100% (470245/470245), done.
Writing objects: 100% (3064514/3064514), 703.02 MiB | 10.84 MiB/s, done.
Total 3064514 (delta 2569775), reused 3059081 (delta 2565342)
To 127.0.0.1:/tmp/repo
* [new branch] master -> master
Is there a remote git config setting which can make the push succeed? Or do I have to repack the repo locally before pushing (with what settings)?
Please note that using a different server with more memory is not an option. Adding memory to the existing server is an option, up to 96MB. It's OK for me to use more disk space than usual on the server if the memory limit is met.
Similar question without a working solution: https://serverfault.com/questions/372899/git-fails-to-push-with-error-out-of-memory
Repacking the repository locally didn't help, git push prints the same error. Repack settings in the local repo:
git config core.packedgitlimit 32m
git config core.packedgitwindowsize=32m
git config pack.threads 1
git config pack.deltacachesize 8m
git config pack.windowmemory 32m
git config pack.packsizelimit 500m
My idea is that the reason why it fails is that the total number of objects is too large: even the SHA-1 hashes won't fit (20 * 3064514 bytes is almost 64MB).
Possible other causes
As #torek pointed out in his comment, this may not be an indication of the server running out of memory, but an indication that something is going wrong locally. Perhaps something changed between when you were pushing to server and to local host that freed up memory on your local machine?
It's also plausible that git is figuring out that you're pushing to localhost, and bypassing the "Git aware" transport mechanism and/or using hardlinks, which might reduce the memory needed. I don't see any indication in the docs that it WOULD do this, and I'm not sure off the top of my head how you could test this, or force it not to do that, but it's a possibility.
Another possible issue is that the host.xz:path/to/repo.git/ url syntax is only recognized if there are no slashes before the first colon, so depending on what server is, that could be causing problems.
If none of these are the case, and the problem is in fact that it's running out of memory on the server, you might have a few options here, depending on the circumstances. I don't know if any of these will work, but they're worth a try.
Solution 1: don't push all the commits at once
I'm assuming you've got many commits in the commit history of master. Try pushing them in stages. E.g.
git push server:/tmp/repo master~500
git push server:/tmp/repo master~400
git push server:/tmp/repo master~300
git push server:/tmp/repo master~200
git push server:/tmp/repo master~100
git push server:/tmp/repo master
Solution 2: Push individual objects one at a time
This is going to be incredibly tedious and DEFINITELY need to be automated/scripted on your local machine. However, you don't actually need to push whole commits all at once.
Instead, you can push individual objects one at a time as long as you push them to a tag ref instead of a branch ref. E.g. if we were working with https://github.com/llvm/llvm-project and wanted to push the tree object 0082ee0b3ad78ff55b2a3a65ef5bfdb8cd9713a1 from it (this is the tree object pointed to by commit faf5e0ec737a676088649d7c13cb50f3f91a703a), we could do git push server:/tmp/repo 0082ee0b3ad78ff55b2a3a65ef5bfdb8cd9713a1:refs/tags/test. Using this we can push individual objects one at a time, starting with blobs, then the tree objects, then finally commit objects. We'd end up with a TON of tags to clean up later, but I'll leave that to you to figure out.
For the rest of these solutions, I'm working under a couple assumptions:
Given the limitations you described, and the way you specified the url as server:/tmp/repo instead of something ending with .git, I'm assuming this remote repository isn't going to be managed with any service like github or gitlab, which should give you a little more room to use some unconventional techniques.
I'm also assuming you probably have the ability to log on to/run
commands on the server.
If either of these are not the case, and the above didn't work, I'm out of ideas at the moment.
Solution 3: backwards push using fetch or clone
There's actually nothing special about a server, it's just another git repository that you can trade commits with. The only difference is that a server is usually hosting what's called a bare repository: it doesn't typically keep a working tree of it's own (in other words, it only keeps the contents of the .git folder).
So, try performing the push in reverse using fetch/clone from server:
Push to a third, intermediate server (let's call it server2). Ideally, one with a lot more performance, like a github hosted repo.
Log onto/ssh into server, and from there, clone the repo into /tmp/repo: git clone --bare git#github.com:path/to/your/repo.git.
I would be surprised if this solved anything on it's own, but it's worth trying and step 1 will still set us up for solution 4 and 5. If by chance it does work, you can tidy up by removing server2 as a remote on server: git remote remove origin, then setting up your remotes on your local machine to point towards server instead of server2.
Solution 4: backwards push, but without fetching all the commits at once
Like solution 3, push to an intermediate server, but this time, instead of using clone and fetching everything all at once, fetch the commits in stages:
Log onto/ssh into server, and from there, initialize /tmp/repo as a bare repo:
cd /tmp/repo
git init --bare
git remote add origin git#github.com:path/to/your/repo.git
Still on server, fetch commits one at a time:
git fetch origin 569d84fe99e63e830ea036598f7fa7a5f9899d7c
git fetch origin 9aaba9d9bb4fc3648a9417820858086b14b6b73e
git fetch origin faf5e0ec737a676088649d7c13cb50f3f91a703a
Solution 5: backwards push, but using partial and/or shallow clones
Instead of fetching individual commits, we can use partial and or shallow clones to restrict how much we are fetching at once. There is a good write-up explaining what those are on the github blog: https://github.blog/2020-12-21-get-up-to-speed-with-partial-clone-and-shallow-clone/. This time, we can't won't use a bare repository. We want to be able to check out commits to fill in the missing objects later. You can follow the instructions here to convert it to a bare repository when you're done. Alternatively, instead of using a regular (non-bare) repository, explicitly fetching the objects might also work, but I don't know for sure off the top of my head.
I think everything I've already written, combined with that write-up should give you all the pieces you need to figure out how to do this. I've already spent hours writing this up, it's late, this solution's kind of complicated, and it's an esoteric question that hasn't been touched in years. If somebody comes across this and needs a more complete answer for this, leave a comment and I'll fill it in, but this is as far as I'm willing to go right now for some potential internet points if nobody actually needs this answer XD.
When I enter this query:
sqlite> DELETE FROM mails WHERE (id = 71);
SQLite returns this error:
SQL error: database is locked
How do I unlock the database so this query will work?
In windows you can try this program http://www.nirsoft.net/utils/opened_files_view.html to find out the process is handling db file. Try closed that program for unlock database
In Linux and macOS you can do something similar, for example, if your locked file is development.db:
$ fuser development.db
This command will show what process is locking the file:
> development.db: 5430
Just kill the process...
kill -9 5430
...And your database will be unlocked.
I caused my sqlite db to become locked by crashing an app during a write. Here is how i fixed it:
echo ".dump" | sqlite old.db | sqlite new.db
Taken from: http://random.kakaopor.hu/how-to-repair-an-sqlite-database
The SQLite wiki DatabaseIsLocked page offers an explanation of this error message. It states, in part, that the source of contention is internal (to the process emitting the error). What this page doesn't explain is how SQLite decides that something in your process holds a lock and what conditions could lead to a false positive.
This error code occurs when you try to do two incompatible things with a database at the same time from the same database connection.
Changes related to file locking introduced in v3 and may be useful for future readers and can be found here: File Locking And Concurrency In SQLite Version 3
If you want to remove a "database is locked" error then follow these steps:
Copy your database file to some other location.
Replace the database with the copied database. This will dereference all processes which were accessing your database file.
Deleting the -journal file sounds like a terrible idea. It's there to allow sqlite to roll back the database to a consistent state after a crash. If you delete it while the database is in an inconsistent state, then you're left with a corrupted database. Citing a page from the sqlite site:
If a crash or power loss does occur and a hot journal is left on the disk, it is essential that the original database file and the hot journal remain on disk with their original names until the database file is opened by another SQLite process and rolled back. [...]
We suspect that a common failure mode for SQLite recovery happens like this: A power failure occurs. After power is restored, a well-meaning user or system administrator begins looking around on the disk for damage. They see their database file named "important.data". This file is perhaps familiar to them. But after the crash, there is also a hot journal named "important.data-journal". The user then deletes the hot journal, thinking that they are helping to cleanup the system. We know of no way to prevent this other than user education.
The rollback is supposed to happen automatically the next time the database is opened, but it will fail if the process can't lock the database. As others have said, one possible reason for this is that another process currently has it open. Another possibility is a stale NFS lock, if the database is on an NFS volume. In that case, a workaround is to replace the database file with a fresh copy that isn't locked on the NFS server (mv database.db original.db; cp original.db database.db). Note that the sqlite FAQ recommends caution regarding concurrent access to databases on NFS volumes, because of buggy implementations of NFS file locking.
I can't explain why deleting a -journal file would let you lock a database that you couldn't before. Is that reproducible?
By the way, the presence of a -journal file doesn't necessarily mean that there was a crash or that there are changes to be rolled back. Sqlite has a few different journal modes, and in PERSIST or TRUNCATE modes it leaves the -journal file in place always, and changes the contents to indicate whether or not there are partial transactions to roll back.
the SQLite db files are just files, so the first step would be to make sure it isn't read-only. The other thing to do is to make sure that you don't have some sort of GUI SQLite DB viewer with the DB open. You could have the DB open in another shell, or your code may have the DB open. Typically you would see this if a different thread, or application such as SQLite Database Browser has the DB open for writing.
My lock was caused by the system crashing and not by a hanging process. To resolve this, I simply renamed the file then copied it back to its original name and location.
Using a Linux shell that would be:
mv mydata.db temp.db
cp temp.db mydata.db
If a process has a lock on an SQLite DB and crashes, the DB stays locked permanently. That's the problem. It's not that some other process has a lock.
I had this problem just now, using an SQLite database on a remote server, stored on an NFS mount. SQLite was unable to obtain a lock after the remote shell session I used had crashed while the database was open.
The recipes for recovery suggested above did not work for me (including the idea to first move and then copy the database back). But after copying it to a non-NFS system, the database became usable and not data appears to have been lost.
Some functions, like INDEX'ing, can take a very long time - and it locks the whole database while it runs. In instances like that, it might not even use the journal file!
So the best/only way to check if your database is locked because a process is ACTIVELY writing to it (and thus you should leave it the hell alone until its completed its operation) is to md5 (or md5sum on some systems) the file twice.
If you get a different checksum, the database is being written, and you really really REALLY don't want to kill -9 that process because you can easily end up with a corrupt table/database if you do.
I'll reiterate, because it's important - the solution is NOT to find the locking program and kill it - it's to find if the database has a write lock for a good reason, and go from there. Sometimes the correct solution is just a coffee break.
The only way to create this locked-but-not-being-written-to situation is if your program runs BEGIN EXCLUSIVE, because it wanted to do some table alterations or something, then for whatever reason never sends an END afterwards, and the process never terminates. All three conditions being met is highly unlikely in any properly-written code, and as such 99 times out of 100 when someone wants to kill -9 their locking process, the locking process is actually locking your database for a good reason. Programmers don't typically add the BEGIN EXCLUSIVE condition unless they really need to, because it prevents concurrency and increases user complaints. SQLite itself only adds it when it really needs to (like when indexing).
Finally, the 'locked' status does not exist INSIDE the file as several answers have stated - it resides in the Operating System's kernel. The process which ran BEGIN EXCLUSIVE has requested from the OS a lock be placed on the file. Even if your exclusive process has crashed, your OS will be able to figure out if it should maintain the file lock or not!! It is not possible to end up with a database which is locked but no process is actively locking it!!
When it comes to seeing which process is locking the file, it's typically better to use lsof rather than fuser (this is a good demonstration of why: https://unix.stackexchange.com/questions/94316/fuser-vs-lsof-to-check-files-in-use). Alternatively if you have DTrace (OSX) you can use iosnoop on the file.
I added "Pooling=true" to connection string and it worked.
This error can be thrown if the file is in a remote folder, like a shared folder. I changed the database to a local directory and it worked perfectly.
I found the documentation of the various states of locking in SQLite to be very helpful. Michael, if you can perform reads but can't perform writes to the database, that means that a process has gotten a RESERVED lock on your database but hasn't executed the write yet. If you're using SQLite3, there's a new lock called PENDING where no more processes are allowed to connect but existing connections can sill perform reads, so if this is the issue you should look at that instead.
I have such problem within the app, which access to SQLite from 2 connections - one was read-only and second for writing and reading. It looks like that read-only connection blocked writing from second connection. Finally, it is turns out that it is required to finalize or, at least, reset prepared statements IMMEDIATELY after use. Until prepared statement is opened, it caused to database was blocked for writing.
DON'T FORGET CALL:
sqlite_reset(xxx);
or
sqlite_finalize(xxx);
I just had something similar happen to me - my web application was able to read from the database, but could not perform any inserts or updates. A reboot of Apache solved the issue at least temporarily.
It'd be nice, however, to be able to track down the root cause.
lsof command on my Linux environment helped me to figure it out that a process was hanging keeping the file open.
Killed the process and problem was solved.
This link solve the problem. : When Sqlite gives : Database locked error
It solved my problem may be useful to you.
And you can use begin transaction and end transaction to not make database locked in future.
Should be a database's internal problem...
For me it has been manifested after trying to browse database with "SQLite manager"...
So, if you can't find another process connect to database and you just can't fix it,
just try this radical solution:
Provide to export your tables (You can use "SQLite manager" on Firefox)
If the migration alter your database scheme delete the last failed migration
Rename your "database.sqlite" file
Execute "rake db:migrate" to make a new working database
Provide to give the right permissions to database for table's importing
Import your backed up tables
Write the new migration
Execute it with "rake db:migrate"
In my experience, this error is caused by: You opened multiple connections.
e.g.:
1 or more sqlitebrowser (GUI)
1 or more electron thread
rails thread
I am nore sure about the details of SQLITE3 how to handle the multiple thread/request, but when I close the sqlitebrowser and electron thread, then rails is running well and won't block any more.
I ran into this same problem on Mac OS X 10.5.7 running Python scripts from a terminal session. Even though I had stopped the scripts and the terminal window was sitting at the command prompt, it would give this error the next time it ran. The solution was to close the terminal window and then open it up again. Doesn't make sense to me, but it worked.
I just had the same error.
After 5 minets google-ing I found that I didun't closed one shell witch were using the db.
Just close it and try again ;)
I had the same problem. Apparently the rollback function seems to overwrite the db file with the journal which is the same as the db file but without the most recent change. I've implemented this in my code below and it's been working fine since then, whereas before my code would just get stuck in the loop as the database stayed locked.
Hope this helps
my python code
##############
#### Defs ####
##############
def conn_exec( connection , cursor , cmd_str ):
done = False
try_count = 0.0
while not done:
try:
cursor.execute( cmd_str )
done = True
except sqlite.IntegrityError:
# Ignore this error because it means the item already exists in the database
done = True
except Exception, error:
if try_count%60.0 == 0.0: # print error every minute
print "\t" , "Error executing command" , cmd_str
print "Message:" , error
if try_count%120.0 == 0.0: # if waited for 2 miutes, roll back
print "Forcing Unlock"
connection.rollback()
time.sleep(0.05)
try_count += 0.05
def conn_comit( connection ):
done = False
try_count = 0.0
while not done:
try:
connection.commit()
done = True
except sqlite.IntegrityError:
# Ignore this error because it means the item already exists in the database
done = True
except Exception, error:
if try_count%60.0 == 0.0: # print error every minute
print "\t" , "Error executing command" , cmd_str
print "Message:" , error
if try_count%120.0 == 0.0: # if waited for 2 miutes, roll back
print "Forcing Unlock"
connection.rollback()
time.sleep(0.05)
try_count += 0.05
##################
#### Run Code ####
##################
connection = sqlite.connect( db_path )
cursor = connection.cursor()
# Create tables if database does not exist
conn_exec( connection , cursor , '''CREATE TABLE IF NOT EXISTS fix (path TEXT PRIMARY KEY);''')
conn_exec( connection , cursor , '''CREATE TABLE IF NOT EXISTS tx (path TEXT PRIMARY KEY);''')
conn_exec( connection , cursor , '''CREATE TABLE IF NOT EXISTS completed (fix DATE, tx DATE);''')
conn_comit( connection )
One common reason for getting this exception is when you are trying to do a write operation while still holding resources for a read operation. For example, if you SELECT from a table, and then try to UPDATE something you've selected without closing your ResultSet first.
I was having "database is locked" errors in a multi-threaded application as well, which appears to be the SQLITE_BUSY result code, and I solved it with setting sqlite3_busy_timeout to something suitably long like 30000.
(On a side-note, how odd that on a 7 year old question nobody found this out already! SQLite really is a peculiar and amazing project...)
Before going down the reboot option, it is worthwhile to see if you can find the user of the sqlite database.
On Linux, one can employ fuser to this end:
$ fuser database.db
$ fuser database.db-journal
In my case I got the following response:
philip 3556 4700 0 10:24 pts/3 00:00:01 /usr/bin/python manage.py shell
Which showed that I had another Python program with pid 3556 (manage.py) using the database.
An old question, with a lot of answers, here's the steps I've recently followed reading the answers above, but in my case the problem was due to cifs resource sharing. This case is not reported previously, so hope it helps someone.
Check no connections are left open in your java code.
Check no other processes are using your SQLite db file with lsof.
Check the user owner of your running jvm process has r/w permissions over the file.
Try to force the lock mode on the connection opening with
final SQLiteConfig config = new SQLiteConfig();
config.setReadOnly(false);
config.setLockingMode(LockingMode.NORMAL);
connection = DriverManager.getConnection(url, config.toProperties());
If your using your SQLite db file over a NFS shared folder, check this point of the SQLite faq, and review your mounting configuration options to make sure your avoiding locks, as described here:
//myserver /mymount cifs username=*****,password=*****,iocharset=utf8,sec=ntlm,file,nolock,file_mode=0700,dir_mode=0700,uid=0500,gid=0500 0 0
I got this error in a scenario a little different from the ones describe here.
The SQLite database rested on a NFS filesystem shared by 3 servers. On 2 of the servers I was able do run queries on the database successfully, on the third one thought I was getting the "database is locked" message.
The thing with this 3rd machine was that it had no space left on /var. Everytime I tried to run a query in ANY SQLite database located in this filesystem I got the "database is locked" message and also this error over the logs:
Aug 8 10:33:38 server01 kernel: lockd: cannot monitor 172.22.84.87
And this one also:
Aug 8 10:33:38 server01 rpc.statd[7430]: Failed to insert: writing /var/lib/nfs/statd/sm/other.server.name.com: No space left on device
Aug 8 10:33:38 server01 rpc.statd[7430]: STAT_FAIL to server01 for SM_MON of 172.22.84.87
After the space situation was handled everything got back to normal.
If you're trying to unlock the Chrome database to view it with SQLite, then just shut down Chrome.
Windows
%userprofile%\Local Settings\Application Data\Google\Chrome\User Data\Default\Web Data
or
%userprofile%\Local Settings\Application Data\Google\Chrome\User Data\Default\Chrome Web Data
Mac
~/Library/Application Support/Google/Chrome/Default/Web Data
From your previous comments you said a -journal file was present.
This could mean that you have opened and (EXCLUSIVE?) transaction and have not yet committed the data. Did your program or some other process leave the -journal behind??
Restarting the sqlite process will look at the journal file and clean up any uncommitted actions and remove the -journal file.
As Seun Osewa has said, sometimes a zombie process will sit in the terminal with a lock aquired, even if you don't think it possible. Your script runs, crashes, and you go back to the prompt, but there's a zombie process spawned somewhere by a library call, and that process has the lock.
Closing the terminal you were in (on OSX) might work. Rebooting will work. You could look for "python" processes (for example) that are not doing anything, and kill them.