-
Notifications
You must be signed in to change notification settings - Fork 667
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
ownCloud client 1.7.1 won't sync files on exfat drive #2693
Comments
@andrewbridge we've been tracking and trying to reproduce this issue in #2701, but without success so far. To get more info, could you try running owncloud with --logfile log.txt again? It's surprising that it should be empty. I'm particularly interested in the lines in the log file around the error message "ERR: Failed to sqlite3 open statedb - bail out:". That could help with figuring out what's going on. |
Hi @ckamm, thanks for taking a look at my issue! I ran owncloud with the --logfile flag as suggested, but nothing seemed to occur. I used the log file panel in the client (F12), and left it for a while, and some output did appear. I've put it in a pastebin, let me know if you'd prefer me to put it elsewhere. Unfortunately, the line you're looking for doesn't appear to be there. Have I not set a verbosity level high enough somewhere? Just to double check, running the client from the command line with the flag should output a file into the same directory, called log.txt? Or have I been looking in the wrong place? Thanks again for your help. |
@andrewbridge The log window has a maximum number of lines, unfortunately. With "--logfile some/path.txt" a logfile should be written (sometimes only fully filled when you close the client). |
@andrewbridge Thank you for the full log! Relevant excerpt:
It looks like we get a "disk I/O error" when running quick_check on the database. |
@andrewbridge If the DB isn't too private and you still have it: could you mail E:\Andrew\Documents/.csync_journal.db* to me (kamm at incasoftware.de)? I want to check whether the DB really is corrupt and whether I can reproduce the problem by opening it on an exfat drive. |
Related to #2697 |
@ckamm I've sent that to the e-mail you listed. As previously noted, I have deleted the db files previously and reinstalled the application in case something (probably me) had corrupted them, and this never seems to help the issue. I'm hoping it's not something that I'm doing wrong! |
@andrewbridge Thank you for the DB files. It looks like a valid empty database to me, I can open it just fine. I currently don't know what's going on. Here's another interesting excerpt from your log file: the db setup:
We create a new database and run all the PRAGMA statements. Then CREATE TABLE fails to exec() and we fail to report the SQL error for that (will fix). The error about committing (triggered by sqlFail()) is also odd, because it looks like there was no error when startTransaction() ran earlier. I'll try to find if similar issues have been reported elsewhere. I'll also fix the error reporting and will ask you to try again once the new builds are done. But probably that error is just a generic "disk i/o error" too. |
OWNCLOUD_SQLITE_JOURNAL_MODE: To use something else than WAL OWNCLOUD_SQLITE_TEMP_STORE: To test with storing temporaries in memory.
@andrewbridge Could I ask you to test this some more?
Please use a new sync dir for each test (or delete the .csync_journal.db* files) to ensure we're getting a fresh journal each time. If you need any help, like more detailed instructions on one of the steps, let me know! I think it's likely that the changed settings will make it work - and if you verify that they do, we'll be able to fix this for good. |
@ckamm Sure! I've done as you instructed, using the build "owncloud-1.8.0.4665-nightly20150131-setup.exe" I've removed some details from the logs, as varying levels of success resulted in the log retaining filenames that I'd rather keep private. Don't think that should affect their usefulness, as the beginning of the path should always be included.
As noted the journal mode change appears to work. How does this affect the running of the client? Would it be safe to run the client in this way as a workaround in the meantime? If so, I'll use it to resume synching as normal. Many thanks, I hope these logs are of some help! |
@andrewbridge Thank you for the efforts! This indicates that @guruz intuition was correct and disabling WAL on FAT, like suggested in #2701 is a good idea. I think we switched to the WAL mode for performance reasons - as far as I know running with journal mode = delete should be fine. |
@andrewbridge It looks to me like temp store = memory also lead to successful sync? Both journal mode=delete and temp store=memory still have a couple of disk I/O errors when running quick check on a readonly database, but the default setting has way more errors. Did you delete the sync journals between tests? (these journal mode flags are sticky, so maybe the temp store memory test did use the DELETE journal mode?) |
@ckamm I closed the command prompt between attempts and as far as I'm aware, I deleted the journals both times. I'm pretty sure the client reported an error with the temp store=memory setting. I don't completely understand the logs, but it appeared not to crawl as deeply with the temp store=memory too. I took those two assumptions and concluded that that method hadn't worked. I noticed the recent commits as well! A final thank you to you and every one else that's been looking and persisting with my issue! I really do appreciate the work and help! |
@andrewbridge You're welcome, it wouldn't have been possible without all the data you provided. Thank you! |
Hi All, Just to report and for records, same issue happens on OS X (Mac), synchronizing folder with compression enabled on NTFS mounted via Tuxera. Using Tuxera NTFS on Mac I've run across problem with compressed folders/drive. Once the problem happens (coupe of times) I/O subsystem is locked and there is no possibility to unmount the volume. The only option is to hard power off system. Would be great if this could get fixed as I'd need to use compression on the given folder. Attached are full logs with excerpt below.
Problem reproduction steps. Tuxera: 2015.3 Volume: NTFS created natively with Windows tools.
complaining about problem with saving database (see logs). At the same time error in System logs is presented:
==> /tmp/owncloud-ntfs-tuxera-compress-error <==
==> /var/log/system.log <==
==> /tmp/owncloud-ntfs-tuxera-compress-error <==
|
Expected behaviour
ownCloud client successfully syncs data from second internal drive to server.
Actual behaviour
ownCloud client attempts to sync data from second internal drive but fails with one of two messages
"Unable to initialise journal"
"CSync failed to load or create the journal file. Make sure you have read and write permissions in the local sync directory."
Steps to reproduce
Server configuration
Operating system: Ubuntu 14.10
Web server: Apache 2.4.7
Database: PostgreSQL
PHP version: 5.5.9
ownCloud version: 6.0.2
Storage backend: Local
Client configuration
Client version: 1.7.1
Operating system: Windows 8.1
OS language: English
Installation path of client: C:\Program Files (x86)\ownCloud
Logs
owncloud --logwindow
orowncloud --logfile log.txt
The log is empty
Nothing related to ownCloud
Large numbers of errors/logs regarding address books and contacts, I'd imagine this if from using the online interface.
Further notes
The text was updated successfully, but these errors were encountered: