Posts by mint@plagu.ee
(DIR) Post #AcOpl44jUER10DEvrc by mint@plagu.ee
2023-12-02T12:55:06.765480Z
0 likes, 0 repeats
And it's done. 77824 bytes in bad sectors, apparently. Now to put a new SSD.
(DIR) Post #AcOrC4ER72tbGEDb4y by mint@plagu.ee
2023-12-02T13:11:09.593384Z
0 likes, 0 repeats
The label on the old SSD that the chink installed reads "Kston" as it's trying to mimic Kingston. Memory itself might be good, but there's something resembling overheating on the Phison controller. Whateva.
(DIR) Post #AcOv1qnabKSAWI2pZg by mint@plagu.ee
2023-12-02T13:54:11.041390Z
0 likes, 2 repeats
Husky_1701525266839_EQZOLY9LHI.…
(DIR) Post #AhWScSjmKQj42oVPsW by mint@plagu.ee
2024-05-03T18:34:38.624299Z
4 likes, 0 repeats
@lauralt @Cookieman @childrapist1488 @lina @Looming @p @NEETzsche @mint Except fchan never federated with anything but itself thanks to its very different activitypub implementation. So much for the open protocol.https://ryona.agency/notice/AdsBj9g6nagzRyIWcyAlso firecucks froze again while I was searching the link, prompting me to killall -9 it, and when I restored the tabs, your post was gone.
(DIR) Post #AhtWtVt2jyrNBgIsLY by mint@plagu.ee
2024-01-26T15:16:46.694228Z
0 likes, 0 repeats
@pernia What?
(DIR) Post #Al0GxObeo5gFDbMwDo by mint@plagu.ee
2024-03-01T17:06:15.000914Z
1 likes, 0 repeats
@mint Plague's on, baby.
(DIR) Post #Al0HnmpRzecPOwnoem by mint@plagu.ee
2024-08-15T21:25:10.299284Z
2 likes, 0 repeats
Been long enough since the last followlist synchronization I didn't even follow leaf here.
(DIR) Post #Al0IoAazjFLx4Dmt8a by mint@plagu.ee
2024-08-15T21:39:00.423265Z
2 likes, 0 repeats
Anyway, bringing plague back up was a pain. I updated the system, and postgres after migration started sparging out and segfaulting. Had to reemerge @world, updating again in the process, and this took days since I'm lazy and so started from scratch every time some package failed to build. In meantime also purged all activities and objects except posts and follows to hopefully make it run faster on hardware that's even worse than agency's server.
(DIR) Post #Al0IoCHDTkikHSY7aS by mint@plagu.ee
2024-08-15T21:39:49.372917Z
0 likes, 0 repeats
Also there's been some issues with clock source in VM that spammed the hell out of dmesg, added tsc=unstable to params and now it seems fine.
(DIR) Post #Al0JyJ2bHgcS6r652G by mint@plagu.ee
2024-08-15T21:47:02.094191Z
1 likes, 0 repeats
Need to synchronize relay list as well.
(DIR) Post #Al2MJAaOVFesC0trWa by mint@plagu.ee
2024-08-16T21:28:45.958685Z
0 likes, 0 repeats
@adachi I think it is, I've put it all the way back in 2022.
(DIR) Post #Al2mv2ikn8SfbzWvFA by mint@plagu.ee
2024-08-17T02:26:57.850101Z
1 likes, 0 repeats
So much faster than agency despite running in a VM with a third of resources and still being subscribed to a shitton of relays. Guess not having any other users makes the difference.
(DIR) Post #Al2nCehmosXELDTZLM by mint@plagu.ee
2024-08-17T02:30:08.796909Z
0 likes, 0 repeats
The DB is 38 gigs on disk compared to agency's 63. Don't know what other differences could be there, likely can't be postgres since accela (from local mirror) loads pretty much instantly as well.
(DIR) Post #Al2ntnNOSMf1gbVYRM by mint@plagu.ee
2024-08-17T02:37:56.602780Z
0 likes, 0 repeats
Could maybe strip 5 million likes/reacts like I did here before without breaking anything, but that would reclaim a couple gigs of space at best.Screenshot_20240817_053659.png
(DIR) Post #Al2oUEd6beUGdVEV3g by mint@plagu.ee
2024-08-17T02:44:31.606439Z
0 likes, 0 repeats
That would still account for 21% of total activites. Fuck it, we'll do it live.
(DIR) Post #Al2rCSfBzTTooKQRMG by mint@plagu.ee
2024-08-17T03:14:50.277364Z
0 likes, 0 repeats
Now I am become death, destroyer of rows.Screenshot_20240817_061624.png
(DIR) Post #Al2rMMAF0dY9Y8fOMK by mint@plagu.ee
2024-08-17T03:16:43.182014Z
0 likes, 0 repeats
(I think local likes/reacts should be excluded in case they couldn't be retracted.)Either way, time for repack, only on one table to not waste too much time.
(DIR) Post #Al3RXHt6qU3Lp2AIIi by mint@plagu.ee
2024-08-17T10:02:04.582061Z
0 likes, 0 repeats
Went to sleep in meanwhile, after repacking DB size is at 57 gigs. Not too shabby since now it's even smaller than salon's.
(DIR) Post #AmXbpWutAqNJSculG4 by mint@plagu.ee
2024-09-30T21:10:55.173350Z
0 likes, 0 repeats
Forgot to pay for this pipe, whoopsies.
(DIR) Post #AmXc9hxdowtnVQp7jM by mint@plagu.ee
2024-09-30T21:14:33.873484Z
0 likes, 0 repeats
@feld @thomcat I'm gonna shill NBD as well. Much simpler protocol than iSCSI thus is easier to setup and has lower overhead, similar to Wireguard vs. OpenVPN. I think Scaleway used it as a main storage on some of their VMs, worked reasonably well.