Post-upgrade migration now on 30+ hrs

Hello

I’ve been running an old version of Atrocore for a while (2.1.26), and am in the process of getting up to date. Our database has approx. half a million products in a folder structure that is maximum four levels deep.

It’s been stuck at rebuilding routes for entity Product for more than 30 hours.

08.09.2026 07:30:09 | Copying post-install & post-delete scripts for modules
08.09.2026 07:30:13 | Copying migration scripts
08.09.2026 07:30:13 | Clearing cache
08.09.2026 07:30:14 | Copying frontend files
08.09.2026 07:30:15 | Run migration Atro 2.2.1
08.09.2026 07:30:15 | Run migration Atro 2.2.2
08.09.2026 07:30:17 | Run migration Atro 2.2.3
08.09.2026 07:30:17 | Run migration Atro 2.2.4
08.09.2026 07:30:17 | Run migration Atro 2.2.5
08.09.2026 07:30:17 | Run migration Atro 2.2.6
08.09.2026 07:30:17 | Run migration Atro 2.2.7
08.09.2026 07:30:17 | Run migration Pim 1.15.22
08.09.2026 07:30:17 | Run migration Atro 2.2.8
08.09.2026 07:30:17 | Run migration Atro 2.2.10
08.09.2026 07:30:17 | Run migration Atro 2.2.9
08.09.2026 07:30:17 | Run migration Atro 2.2.11
08.09.2026 07:30:17 | Run migration Atro 2.2.12
08.09.2026 07:30:17 | Run migration Atro 2.2.13
08.09.2026 07:30:17 | Run migration Atro 2.2.14
08.09.2026 07:30:18 | Run migration Atro 2.2.15
08.09.2026 07:30:18 | Run migration Pim 1.15.23
08.09.2026 07:30:19 | Run migration Import 1.10.13
08.09.2026 07:30:19 | Run migration Export 1.10.13
08.09.2026 07:30:19 | Run migration Pim 1.15.24
08.09.2026 07:30:19 | Run migration Pim 1.15.25
08.09.2026 07:30:20 | Run migration Atro 2.2.16
08.09.2026 07:30:21 | Run migration Atro 2.2.17
08.09.2026 07:30:22 | Run migration Atro 2.2.18
08.09.2026 07:30:22 | Run migration Atro 2.2.19
08.09.2026 07:30:22 | Run migration Atro 2.2.20
08.09.2026 07:30:22 | Run migration Atro 2.2.21
08.09.2026 07:30:22 | Run migration Atro 2.2.22
08.09.2026 07:30:22 | Run migration Atro 2.2.23
Rebuilding routes for entity Folder...
Rebuilding routes for entity Product...

It seems like there is activity, though, as the postgres process is churning away at approx. 100% CPU and there is real postgres activity:

  pid  | state  |    duration     |                                                query                                                 
-------+--------+-----------------+------------------------------------------------------------------------------------------------------
 18857 | active | 00:00:00.444955 | UPDATE product SET routes = $1 WHERE routes LIKE $2 OR id = $3
 28003 | active | 00:00:00        | SELECT pid, state, now() - query_start AS duration, left(query, 100) AS query FROM pg_stat_activity 
 27981 | active | 00:00:00.326274 | UPDATE product SET routes = $1 WHERE routes LIKE $2 OR id = $3
(3 rows)

Still… 30+ hrs for a database this size seems like a lot. Any idea why this is taking so long?

Hello @ast
We will investigate the issue

Thanks for investigating! Just to follow up, migration finished after approximately 50 hours. I’m no expert, but I believe the SQL query used in this point is perhaps not as efficient as it could be.