9.2.1.6-RC-71b05dd-x64 TOTAL DISASTER

[panz]

Yes, today I upgraded from 9.2.1.3 to v. 9.2.1.6-RC.

After the upgrade, I started my usual Friday backup. To put it short, it ended with a complete loss of data on my backup Volume!

I have posted one of the bugs
https://bugs.freenas.org/issues/5280

and here are some examples of the errors I'm getting

Jun 20 21:54:40 NAS common.pipesubr: cannot hold snapshot 'storage@auto-20140620.1945-2w': tag already exists on this dataset

Jun 20 22:16:22 NAS notifier: cannot release hold from snapshot 'storage/scripts@auto-20140620.1945-2w': no such tag on this dataset

I'm tired and frustrated :(

 

[cyberjock]

Yikes. I don't do replication on my 9.2.1.6 box but this does still suck. :(

 

[panz]

9.2.1.3 was working like a charm. I hadn't the need to do backups for a while so I didn't use the Replication task for a while (but I did snapshots every night).

Yesterday I fed the server with a bunch of new data so I triggered the Snapshot task and BANG! It killed my destination backup Volume! (my replication task is very simple: the "storage" Volume is recursively replicated to the "backup" Volume; each of one is a RAIDZ2 pool).

So, today I decided to upgrade to v. 9.2.1.6-RC hoping it would have handled this task better... WHAT A MESS! not only it doesn't start the replication after the first successful one, but there are a lot of missing directories!

Reverting to 9.2.1.3 NOW!

 

[Yatti420]

GUI upgrade from 9.2.1.5 to 9.2.1.6 went smooth.. No issues so far..

 

[panz]

GUI upgrade from 9.2.1.5 to 9.2.1.6 went smooth.. No issues so far..

Did you try a Replication?

 

[toadman]

I can confirm replication worked for me from a 9.2.1.6-RC system to another 9.2.1.6-RC system.

 

[panz]

Reinstalled 9.2.1.6 from scratch: first replication = Ok. Second one: wiped all my data.

I'm DESPERATE :(

 

[jkh]

Less emotion, more data please! Seriously, all this shrieking and running around in panic does not bring us one step closer to figuring out what is going on with your replication scenario. Only a clear understanding of your configuration will do that. You should attach your database from the replication system to your ticket.

We have 9.2.1.6 (BETA and RC) replicating back and forth in lots of scenarios right now. Works great. We've also been using it in production, with replication, just fine as well. In fact, with the zfs hold support, it's now no longer possible to reproduce some of the earlier replication problems. It works better for us than ever before, but clearly not for you, so again - we need to figure out what specifically it is YOU are doing to make it fail like this. We need information on how you set up your replication, with *all* of the options you checked, described step by step. Let's take this out of the forums, where no one can help you, and into the bug reporting system where developers can look at the data. Thanks.

 

[panz]

I apologize for my behavior. I'm going to post all the data to this ticket:

https://bugs.freenas.org/issues/5293

I didn't understand when you said: «You should attach your database from the replication system to your ticket».

Which data is needed?

 

[cyberjock]

I *think* he wants you to do a "save config" from the WebGUI and upload the file. BUT... your username and passwords are in the file, so you should change your passwords before uploading the file(or change them after). The file will be available for the world... so do keep that in mind. ;)

 

[panz]

And I suppose that SSH keys are NOT in the database, right?

 

[cyberjock]

They are, but the public key doesn't do anyone any good. You could paste it in the forums and it won't help anyone.

The private key for your server is in there though.. so be cautious. Frankly, I'm somewhat surprised that jkh asked you to post your config file considering the sensitivity of the data in the config file...

 

[joeschmuck]

Thankfully it didn't take out your storage pool. I'll be watching your bug report and hope the developers can reproduce it easily so they can correct the issue.

 

[panz]

Thankfully it didn't take out your storage pool. I'll be watching your bug report and hope the developers can reproduce it easily so they can correct the issue.

A friend of mine gave me his server so, now, I'm copying all my data via Samba to it :)

 

[joeschmuck]

A friend of mine gave me his server so, now, I'm copying all my data via Samba to it :)

Good friend!

 

[panz]

I'm lucky: I have a lot of friends here in the Forum too ;)

 

[DaPlumber]

Just spitballing here, but do you have manual as well as automatic snapshots?

 

[panz]

No, only automatic.

 

[DaPlumber]

No, only automatic.

Then forget I spoke. :sympathy: either way...

 

[panz]

You gave me an idea: one of these days I'll try a manual snapshot + manual send/receive :)