Has macOS Sequoia 15.2 broken ABB

[Bill Wilder]

UPDATE: Have seen the failure on a third Mac. This posting from the SuperDuper folks points to a macOS bug: Shirt Pocket Watch - You’re a Mean One

Please consider this preliminary - I am still troubleshooting. ABB backup jobs for two Macs here have started failing coincident with those systems upgrading to macOS Sequoia 15.2 (from 15.1). DSS, ABB Package and ABB Agent fully up-to-date. Jobs now taking a very long time, GB transferred increases but percentage done ceases to update around 70 or 80 percent done and some time later backup fails with error -3 (Failed to run asr).

This might be local to my circumstances but wanted to post this heads up in case anyone is observing failures (or continued success) post 15.2

 

[Rusty]

I can confirm that my m4 mini with 15.2 has started to fail ABB backups. Just seen the notifications for the past two days but haven’t looked into it yet as I didn’t had time

Other Macs that are still not on 15.2 are running fine.

 

[fredbert]

My Mini M1's ABB task failed after the macOS 15.2 update. The task I run is once a day at 22:00 and the macOS update was during the day on Thursday, so the update didn't interrupt a running task. The failed backup alert was Friday morning (or that was when I saw it). I didn't have time to investigate so ABB server, task, and Mac haven't changed state yet.

Looking this morning (Saturday) I see last night's backup task is still running now, 10 and a half hours later (08:30), and looks to be doing the usual thing after a failed run... it's doing a full transfer from Mac to NAS even if it doesn't actually change much in the stored NAS data. The task is 88% complete as of now; the network traffic between ABB agent and NAS is still high for the 1 GbE NAS port (80-100 MB/s).

For now I would leave it and see if it completes. If it fails again the usual next step I take is to log out the ABB agent (needs a DSM admin user) and then sign in again (using my normal DSM user). This initiates a full backup run too.

 

[Rusty]

I have just checked and the status was Error -3, classic ASR.

How do I fix Active Backup for Business error codes -1, -2, and -3? - Synology Knowledge Center

Synology Knowledge Center offers comprehensive support, providing answers to frequently asked questions, troubleshooting steps, software tutorials, and all the technical documentation you may need.

kb.synology.com

I have added the exceptions listed, and rerun it. I will see if it will pass.

UPDATE:

I have deleted the last 2 failed backup versions from the list and run the backup manually from the ABB console.

 

[Rusty]

Backup failed in the end again with the same error. It took 1,5 +/- with major gaps (nothing odd about ABBs method of backup).

Looking into abb logs currently to see if I can detect are their any files that are causing this specific issue.

These look like repeated attempts, so it could be that indeed something is causing issues at a specific segment on the drive and or OS.

 

[Rusty]

This time around I watched the progress actively. The gaps were much shorter, as the Mac was not in sleep mode, but the pattern is the same.

So deleting the failed job, and rebooting the Mac again did not help and ABB behaved as follows:

Job started at 13:22 and attempted to do a delta update 5 times in the next 10 min. At 13:33 (yellow line) and 88% the same job reported that it cannot run ASR, code "-3", but the same task continued registering that the ABB will perform full backup.

What follows is the same pattern again, 5 times. Two small peaks (macOS main volume) and a big data volume block. After the same pattern of 11GB + 60GB, the process repeats. After the 5th attempt, the task registered as failed with the before mentioned error -3.

 

[Bill Wilder]

This time around I watched the progress actively. The gaps were much shorter, as the Mac was not in sleep mode, but the pattern is the same.

So deleting the failed job, and rebooting the Mac again did not help and ABB behaved as follows:

View attachment 21287

Job started at 13:22 and attempted to do a delta update 5 times in the next 10 min. At 13:33 (yellow line) and 88% the same job reported that it cannot run ASR, code "-3", but the same task continued registering that the ABB will perform full backup.

What follows is the same pattern again, 5 times. Two small peaks (macOS main volume) and a big data volume block. After the same pattern of 11GB + 60GB, the process repeats. After the 5th attempt, the task registered as failed with the before mentioned error -3.

I suspect we may need to wait a macOS update for this fix. Did you see the Edit I applied to my OP (re: SuperDuper)? Or perhaps Synology can work around this with an agent update. For now I've suspended my ABB jobs and will fall back to Time Machine.

 

[Rusty]

I suspect we may need to wait a macOS update for this fix. Did you see the Edit I applied to my OP (re: SuperDuper)? Or perhaps Synology can work around this with an agent update. For now I've suspended my ABB jobs and will fall back to Time Machine.

I have seen your edit yes, just wanted to confirm the behavior. This will be much easier for Syno to fix by alerting Apple, then us ofc, so the more cases are reported the sooner this will be fixed. I have updated only my test M4 atm that is of no production value to me, so will wait for a confirmed update before applying it to my other machines.

Imho, Syno will wait for apple to patch 15.2 before they will implement their workaround (if any), just because, this is a macOS "bug". If it turns out that Apple has set a new direction moving forward in terms of the replicator mechanism, then all 3rd parties will have to adjust. In any event, while this does look like a bug, there is nothing more to do but to wait.

I have opened the case with Syno on this matter and reported the issue.

 

[Bill Wilder]

I have seen your edit yes, just wanted to confirm the behavior. This will be much easier for Syno to fix by alerting Apple, then us ofc, so the more cases are reported the soon this will be fixed. I have updated only my test M4 atm that is of no production value to me, so will wait for a confirmed update before applying it to my other machines.

Imho, Syno will wait for apple to patch 15.2 before they will implement their workaround (if any), just because, this is a macOS "bug". If it turns out that Apple has set a new direction moving forward in terms of the replicator mechanism, them all 3rd parties will have to adjust. In any event, while this does look like a bug, there is nothing more to do but to wait.

I have opened the case with Syno on this matter and reported the issue.

Thanks for notifying Synology. Appreciate it.

 

[Rusty]

Support reply

Thank you for contacting Synology Support. We appreciate your patience and understanding.

We have collected the system log and escalated it to our developers for further analysis. They will investigate the issue thoroughly to identify the root cause.

Rest assured that we are actively working on a solution for you. We will keep you updated on the progress and provide any relevant information as soon as it becomes available.

 

[fredbert]

I can confirm that I haven't been able to complete a backup too. It's tried three times now, and at least two where full backup attempts. I'll see what CCC does tonight on the weekly backup. I hate having to rely on Time Machine as it's never really been that reliable for me.

 

[Koll77]

Confirm also (after trying all Synology recommended "fixes" for error code -3) that I cannot successfully complete ABB. Every attempt gets to about 80% then it hits a "snag" and starts over again eventually failing after about 1.5 to 2 hours. Suspect as others that an incompatibility exists. Sequoia 15.2 (worked with earlier version).

 

[Telos]

ABB is/has been/will always be crap. Even when the backup runs, it is only a security placebo unless you do a full restore. But few attempt so, until the sad day they discover they cannot. Buggy and slow. Good luck with any patch.

 

[Rusty]

Another Synology reply

We appreciate your report and our developer has noticed this issue.

After investigation, similar issues with products related to using ASR (such as CCC or SuperDuper) have the same issue.

We will continue to monitor whether macOS has any related issue fixes. Any updates will be shared as soon as they are available.

Safe to say, they will not be implementing a fix, but rather we will have to wait for a macOS patch.

 

[fredbert]

My weekly CCC task ran successfully overnight, with around +57 GB and -45 Gb of changes . It's a clone of the data partition, with snapshots and SafetyNet. The destination is an external USB HDD drive. I can't see where it is different from other backups that rely of macOS Migration Assistant.

ABB still trying and failing. I'll disconnect the Mac from ABB server and then try re-adding it. Bound to fail afterwards, but hey!

Update (+ a few minutes): I remember sometime before it was necessary to uninstall the ABB agent (go to click-click the ABB client app in /Applications and you'll find the uninstaller in the Contents folder), so did this too. During the uninstall the Mac crashed, which might be for any reason: I logged the agent out while it was doing a backup; the changes in macOS had some inconsistency with the ABB files; it just felt like it, being Monday morning, and all that entails.

Now have the ABB client app and agent reinstalled, connected, and task running. I expect it to fail, but will report back either way.

 

[fredbert]

Now have the ABB client app and agent reinstalled, connected, and task running. I expect it to fail, but will report back either way.

I'm available for fortune-telling, reasonable rates, accuracy guaranteed.

It failed.

 

[fredbert]

It failed.

Ho and hum. What failed would seem to have been it trying to complete the previous cancelled and pre-logged out task.

A new run just completed successfully. I did, as @Rusty was saying to do, delete previous failed attempts in the Version list.

 

[Telos]

WOW. 3 successive posts... for a moment I thought you must be possessed by J

 

[Rusty]

Ho and hum. What failed would seem to have been it trying to complete the previous cancelled and pre-logged out task.

A new run just completed successfully. I did, as @Rusty was saying to do, delete previous failed attempts in the Version list.

Hmm I have deleted all failed backups each time but the case was as explained. It still runs just fine on other systems and full recovery is also functional just fine. It’s only the issue with 15.2 for me.

Will look into it again tomorrow but apart from destroying other versions, not sure what else there is to do right now.

The fact that it tried to do a full backup 5 times and failed each time exactly the same is telling me that there is some sort of bug.

Still, glad you got it going again, I do wonder if the increments will continue without issue as well.

 

[fredbert]

WOW. 3 successive posts... for a moment I thought you must be possessed by J

Yes. Hadn’t noticed it was three. I could’ve edited them together. Sorry.

Back to ABB and macOS 15.2, I think uninstalling and reinstalling the ABB app/client might have been why it worked. But then it also could have been the Mac’s reboot (even though from a crash). Or both events. When I get the chance to try this on the other Mac M1 it may shed some light on this.