ERROR: Need for reconciliation detected! Please run the reconcile task...

[onlinespending]

I recently replied to a thread in the WHS sub-forum, but realize it's sort of out of place. Somehow my FlexRAID Storage Pool (running T1+ Real-Time RAID4) was missing one time i restarted my computer and when I tried to manually start the Storage Pool I received this "ERROR: Need for reconciliation detected! Please run the reconcile task... ". So I went ahead and reconciled the pool and everything was fine. My FlexRAID pool started up again and all data was there. Problem is, when I restarted after that the same issue occurred. It's telling me that I need to reconcile again! Seems like this will be a reoccurring problem, needing to reconcile on every reboot.

Now that I look at the log a bit, it looks as if it also says there are some "uncommitted pending deletes detected". Not sure if I needed to take care of those after the first reconciliation, which may have lead to it requiring a reconciliation after rebooting. I'm a bit confused how that even happened in the first place, but is there a way to avoid having to reconcile on a reboot?

Code: [Select]

[2012-03-14 15:52:03,422] INFO : RealTime RAID reconciliation success!
[2012-03-14 15:52:03,422] WARN : Uncommitted pending deletes detected: please run the Forced Synch Verify task!
[2012-03-14 15:52:03,422] WARN : Uncommitted pending deletes detected: please run the Forced Synch Verify task!
[2012-03-14 15:52:06,308] INFO : Backing up Real-Time RAID filesystem DBs to C:\Program Files\OpenEgg.org\FlexRAID-Backups
[2012-03-14 15:52:06,324] INFO : Executing copy-live-dbs...
[2012-03-14 15:52:06,324] INFO : [copy-live-dbs] started at: Wed Mar 14 15:52:06 EDT 2012
[2012-03-14 15:52:06,371] INFO : [copy-live-dbs] completed at: Wed Mar 14 15:52:06 EDT 2012
[2012-03-14 16:29:31,234] INFO : Initiating FlexRAID Storage Pool...
[2012-03-14 16:29:31,265] INFO : Host OS = Microsoft Windows [Version 6.1.7601]
[2012-03-14 16:29:31,265] INFO : Building View Registry...
[2012-03-14 16:29:31,265] INFO : Building Auto View Registry...
[2012-03-14 16:29:31,296] INFO : Loading View options...
[2012-03-14 16:29:31,749] INFO : Using Auto-View definition...
[2012-03-14 16:29:31,749] INFO : Verifying real-time RAID previous exit state...
[2012-03-14 16:29:31,749] WARN : Need to run the Forced Sync Verify task detected! The pool will mount in read-only mode...
[2012-03-14 16:29:31,749] WARN : Please run the Forced Sync Verify task to fix the storage pool....
[2012-03-14 16:29:31,749] WARN : Disabling all live operations! All future operations will get an access denied error...
[2012-03-14 16:29:31,749] INFO : Building View Handler...
[2012-03-14 16:29:31,905] INFO : Mounting Storage Pool drive...

[onlinespending]

OK, so I figured I had discovered the issue, but apparently I'm wrong as I'm still getting this error. Keep in mind I just installed FlexRAID 2 days ago, so to be consistently having these issues isn't exactly reassuring.

My setup is as follows. I have an SSD that I've partitioned as C: for Windows and D: for some files. I have four 2TB drives that I had originally as E:, F:, G:, H: respectively. I then created the FlexRAID RT RAID using these four drives. But then I shutdown the PC and connected another drive directly to a SATA port of the motherboard in order to transfer some backed up files I had to the FlexRAID RT RAID. These drive was assigned E: by Windows. I started the transfer from this backup drive to the FlexRAID RT RAID and somewhere in the middle of the transfer it just stops and the backup drive on E: just disappears. And when I restart the PC FlexRAID issues the reconcile error. So it seemed clear to me that FlexRAID was somehow disturbed by the fact that one of the RAID drives was originally on E: the same drive letter assigned to a drive I connected to the PC after the fact. This sort of explained why the transfer was aborted mid-transfer and the drive disappeared in My Computer.

So I started from scratch but started by first assigning the drives that would be used in the FlexRAID RT RAID as R:, S:, T:, and U: respectively instead. When I went to connect my backup drive to transfer the files to the RAID it was again assigned E:, but this time the transfer completed just fine and the drive did not disappear. So I figured I had discovered a bug with RT RAID. But after repeating this same procedure with another drive (I have several smaller drives with files on them that I wish to transfer to the RAID) and restarting the computer it again complained of requiring a reconciliation. There was no issue with the transfer aborting or the backup drive disappearing, so this is certainly odd.

The only thing I can think of that may be "exceptional" with my setup is that I use a HighPoint RAID controller simply for the SATA ports. The four drives connected to it are in "legacy mode" and are in no way being used in a hardware RAID by the controller. I do have PUIS (powerup in standby) enabled on the drives and staggered spinup enabled with the controller. The moments that I restart the PC (after which FlexRAID issues the reconciliation error) some of the drives may be spundown. I'm not sure if it's required that all drives that are part of a FlexRAID storage pool be actively spinning when the computer is shutdown or restarted. Though this would not be unique to my setup. Anyone could have this issue.

[Brahim]

@onlinespending

RT RAID has a number of issues that I am aware of that I will be fixing in the coming week.

1. The drive letter assigned to the backup drive is really a non-issue and it disappearing was not related to FlexRAID.
Once FlexRAID release a drive letter (by taking ownership of a drive), it really no longer cares about that drive letter until it is time to release the managed drive back to the user. And that only happens when you delete the configuration.

2. I am working on making RT RAID more robust. File deletes in RT RAID are transactional and asynchronous. But I think there is a bug somewhere in the transaction management that is leading to partial transactions and leading to errors that in turn trigger reconciliation need.
I am going to spend next week revamping that whole aspect for the next release.
For now, I recommend that user manually stop the storage pool before restarting or shutting down their PCs as they will give FlexRAID enough time to do some necessary clean ups.

[NLS]

There is no point to make the system heavier by adding code to handle and process logs even big.
If you just implement a way to easily access those logs (for download) right from the web page, it will be more than enough. And while you are at it, same for config files.
Good both for the newbie (no searching where the logs are) and for the expert. This way remotely someone could do proper debugging (or you).
Autozipping (prefer 7z library of course - way best) would be nice.

Of course I need RT resliliency WAY WAY more than anything else (and it's also #1 right now in votes).

[Opterix]

@onlinespending

RT RAID has a number of issues that I am aware of that I will be fixing in the coming week.

1. The drive letter assigned to the backup drive is really a non-issue and it disappearing was not related to FlexRAID.
Once FlexRAID release a drive letter (by taking ownership of a drive), it really no longer cares about that drive letter until it is time to release the managed drive back to the user. And that only happens when you delete the configuration.

2. I am working on making RT RAID more robust. File deletes in RT RAID are transactional and asynchronous. But I think there is a bug somewhere in the transaction management that is leading to partial transactions and leading to errors that in turn trigger reconciliation need.
I am going to spend next week revamping that whole aspect for the next release.
For now, I recommend that user manually stop the storage pool before restarting or shutting down their PCs as they will give FlexRAID enough time to do some necessary clean ups.

Is there any word on improvements to real time raid as I keep getting the same issue.   The pool will fail to start after a restart of the server and it's making my server unusable.

[Spritzup]

Same here.  At this point, the system is unusable... some feedback would be appreciated.

[BiiRO]

Whats happened to this code fix back in March about trying to address this error as its still there. Nearly everytime I reboot my server i have this error..

[Brahim]

Whats happened to this code fix back in March about trying to address this error as its still there. Nearly everytime I reboot my server i have this error..

Make sure you are following the DOs and DON'Ts posted on the wiki.
In all cases, post your logs if you have them in TRACE mode.