On 01/12/2022 19:31, Mattia Rizzolo wrote:
On Wed, Nov 30, 2022 at 06:46:34AM +0100, Andreas Tille wrote:I realise that it is lagging behind what UDD has.[…]Could you please have a look?Already looking at https://udd-mirror.debian.net/ shows that the last update was on the 26th. It seems one import failed due to ENOSPC, despite more than 30 GB available... No idea what's that about, so I just kicked it for now.
This is familiar...On another (non-public) UDD mirror I had a similar problem when ullmann was upgraded to bullseye while the mirror was still running (*cough*) jessie. The format of the postgresql dump changed slightly and so importing the dump into the mirror started failing. Delightfully, pg felt the need to include the entirety of the failed dump in its logs, so the log files grew by many GB every cron run, and ENOSPC ensued, rotating logs cleaned up some space, people were confused.
That said, the timing of this failure doesn't seem right for this to be the problem, which should have hit in the last days of October 2022.
If, however, this is indeed the same problem, then upgrading the mirror so that it has a newer postgresql should be enough. Alternatively, removing a single offending line from the dump makes the dump compatible with the older postgresql:
$ xzcat $DUMP | sed "/SET default_table_access_method = heap;/d" | psql… ymmv! cheers Stuart -- Stuart Prescott http://www.nanonanonano.net/ stuart@nanonanonano.net Debian Developer http://www.debian.org/ stuart@debian.org GPG fingerprint 90E2 D2C1 AD14 6A1B 7EBB 891D BBC1 7EBB 1396 F2F7