Database Upgrade Stalled 1.3.0

My db is on a separate drive because the first v2 upgrade attempt filled up my 100GB OS drive.

I moved the V1 db folder to G: and yaml is pointing to the new location fine. Chia boots and runs.

I just ran the db CLI updater to complete the 1.3.0 upgrade (Windows full farm client)

It commenced. But is frozen on 28%, and has been running for 12 hours.

What do I do now?

Did you run out of disk space on g: drive as well?

Plenty of space left

How much space left you have on your c: drive? Maybe chia is using windows temp folder (usually your c:\windows\temp), and just choked that one? Kind of unlikely, but worth to check.

Also, could you check your g: drive, how big is right now v2 db? Is it there are all?

Lastly, instead of letting chia decide what are the paths / destinations, maybe you could do it explicitly - How should I upgrade Chia v1 DB with boot drive almost out of space? - #3 by Jacek
Just change there that %userprofile%… with g:\path\v1 and g:\path\v2. Maybe to not get chia confused just create another folder for v2.

Lastly, what is your g: drive? is that HD or SSD/NVMe? When I did sqlite vacuum before, it took me roughly 30 mins (on NVMe), so let say double that maybe to get how fast it should go. Of course a bit more for HD. Still, not 12 hours for 30%. With that speed you can just scratch your db, and sync from zero.

Chia does not leave the source db unchanged while it creates the compressed db?
Once you start the compression process, it is all or nothing?

It should leave it unchanged, so you can go back if the v2 is not working. Although, seeing what is going on with v1.3, I would make a copy before doing the upgrade.

You can look at that upgrade as a straightforward file copy, so I really don’t understand where is the room to mess it up. First it choked his c: drive (making a massive copy and not checking the free space), now it has problems with non-default location. Or maybe it is Monday.

How much space left you have on your c: drive

  • 60GB free

g: drive

  • 100GB free


  • 0kb!

Lastly, what is your g: drive?

  • HDD

So in spite of getting to 30% v2 is 0kb. Something not quite right.

I’m not fully synced… does that matter? Farm has been offline for a week, and it’s taking a while to catch up 1692747/1696153. Even with the port forwarded it’s syncing slower than I’d expect. 48hrs. Port shows as closed on can you see me (it’s open).

Ah crap I closed powershell and Chia and now the V2 DB file is huge

My brother did the upgrade over the weekend, it does not alter the V1 database. It only creates the V2 database up to the synced height when you started the process.

If it’s still syncing a lot of data, that may well slow the process down, especially on a HDD, how much I’ve no idea.

Yeah, when file is still being open for writing (not closed since the beginning), file explorer is not reporting the size.

The problem with that upgrade is that it is reading/writing in rather very small chunks. Therefore drive performance goes down to drain. You may want to restart it the way I posted above (specifying src and dst files), where v1 and v2 will be on different drives.

You may want to monitor disk activity with Task Manager. Most likely, your disk will show 100% utilization, but really very low transfer rates (potentially in a single digit MBps). Also, most likely, you will see there extra *wal file(s). The one for v2 should be changing really fast (good) to indicate v2 writing.

@drhicom How long did it take you to upgrade to v2?

Thanks. Well somehow I lost me v1 files. May have been user error. I distinctly only remember deleting v2 files in order to start again. But somehow V1 is gone and I’m syncing from 0.

I took the opportunity to backup my keys remove everything and reinstall. Defaulting to V2 and 1/3 blockchain synced in 12 hours.

See you in a couple of days :confused:

It’s been emotional.

1 Like

Backups , backups, thats why I let my harvester sync up once in awhile to get a database file other than whats on my USB drive.