GroundSeg ship stuck in pack/meld w/o rebuild option

[sigilante]

Log is pages of:

2025-10-20T13:40:15.092227546Z palloc: off-heap: used: MB/13.210.912
2025-10-20T13:40:15.092254874Z palloc: off-heap: total: MB/15.727.784
2025-10-20T13:40:15.093133055Z mars: pack: gained: MB/5.636.096
2025-10-20T13:40:56.838057858Z palloc: off-heap: used: MB/13.262.088
2025-10-20T13:40:56.838087298Z palloc: off-heap: total: MB/15.727.952
2025-10-20T13:40:56.838929925Z mars: pack: gained: MB/4.751.360
2025-10-20T13:41:44.318122614Z ames: packet dropped (6000 total)
2025-10-20T13:42:19.625512725Z palloc: off-heap: used: MB/13.046.216
2025-10-20T13:42:19.625539689Z palloc: off-heap: total: MB/16.775.844
2025-10-20T13:42:19.626404680Z mars: pack: gained: MB/5.914.624

No Rebuild is available in the UI:

Next step?

[sigilante]

Very strange. Still packing and melding every few seconds but in between it seems to work.

[sigilante]

Restarted everything.

Closing for now but will re-open if it recurs.

[sigilante]

This is recurring in two ships now. They run normally with .run, but a Rebuild in GS does not alter this behavior. One shows a negative time to next pack, but the other shows a normal positive time.

"2025-10-27T15:58:20.241292832Z URTH PACK"
"2025-10-27T15:58:20.242980599Z Running with no STDIN"
"2025-10-27T15:58:20.307292788Z disk: loaded epoch 0i633905001"
"2025-10-27T15:58:20.308258565Z loom: mapped 8192MB"
"2025-10-27T15:58:20.308282800Z boot: protected loom"
"2025-10-27T15:58:20.308294183Z live: mapped: GB/1.042.382.848"
"2025-10-27T15:58:20.344207335Z boot: installed 1770 jets"
"2025-10-27T15:58:20.396299364Z disk: rolling to epoch 634196342"
"2025-10-27T15:59:00.250843554Z loom: image backup complete"
"2025-10-27T15:59:00.667228121Z disk: created epoch 634196342"
"2025-10-27T15:59:03.005971266Z palloc: off-heap: used: MB/6.650.988"
"2025-10-27T15:59:03.005995852Z palloc: off-heap: total: MB/8.385.676"
"2025-10-27T15:59:03.006452637Z urbit: pack: gained: MB/6.193.152"

[sigilante]

I wonder if what it's doing is packing then hitting a log replay that is too slow to show up in the GS logs?

[sigilante]

Looks like GS was stuck packing, then I did a manual replay, then rebuilt the GS, then it packed once again, then it's running. (On ~lagrev-nocfep)

[sigilante]

~magbel is still stuck on packing even with the same cycle.

[sigilante]

Seeing this on a ship now:

2025-11-05T21:06:17.197314316Z URTH PACK
      
2025-11-05T21:06:17.199554540Z Running with no STDIN
      
2025-11-05T21:06:17.246200842Z disk: loaded epoch 0i40837519
      
2025-11-05T21:06:17.246341570Z loom: mapped 2048MB
      
2025-11-05T21:06:17.246493563Z boot: protected loom
      
2025-11-05T21:06:17.246503758Z live: mapped: MB/677.806.080
      
2025-11-05T21:06:17.254054369Z boot: installed 1770 jets
      
2025-11-05T21:06:18.321232188Z palloc: off-heap: used: MB/3.815.972
      
2025-11-05T21:06:18.321251523Z palloc: off-heap: total: MB/5.766.060
      
2025-11-05T21:06:18.321523030Z urbit: pack: gained: KB/49.152
      
2025-11-05T21:06:42.758554804Z /urbit/start_urbit.sh: line 95: n: command not found

[sigilante]

This appears to have been resolved with the latest Groundseg container release (Vere 4.0).

[sigilante]

[sigilante]

Rebuild, pack and meld, didn't work. Finally a chop got it to kick over.

[sigilante]

ha ha NOPE, it's still attempting a new pack and meld on the hour