Conversation
Notices
-
Embed this notice
anime graf mays ?️? (graf@poa.st)'s status on Friday, 06-Sep-2024 22:24:45 JST anime graf mays ?️? @sun how many federator_incoming oban jobs do you have for misskey? -
Embed this notice
Christmas Sun (sun@shitposter.world)'s status on Friday, 06-Sep-2024 22:24:44 JST Christmas Sun @graf I reject misskey, will that make a difference? -
Embed this notice
Christmas Sun (sun@shitposter.world)'s status on Friday, 06-Sep-2024 22:25:13 JST Christmas Sun @graf wait I mean misskey.io. what are we looking for? -
Embed this notice
feld (feld@friedcheese.us)'s status on Friday, 06-Sep-2024 22:25:43 JST feld @sun @graf
@zop was saying in IRC he got flooded with thousands and thousands of jobs from MisskeyIn conversation permalink -
Embed this notice
Christmas Sun (sun@shitposter.world)'s status on Friday, 06-Sep-2024 22:26:32 JST Christmas Sun @graf @feld still confused are we talking about any misskey or just misskey.io In conversation permalink Attachments
-
Embed this notice
anime graf mays ?️? (graf@poa.st)'s status on Friday, 06-Sep-2024 22:26:33 JST anime graf mays ?️? @feld @zop @sun yeah definitely. poast is blocked by misskey.io afaik and we don't have an issue, but my private instance has the same bug with incoming oban jobs from misskey as vd15 yesterday afternoon . 93 million oban jobs for a handful of misskey users In conversation permalink Attachments
-
Embed this notice
feld (feld@friedcheese.us)'s status on Friday, 06-Sep-2024 22:26:36 JST feld @graf @zop @sun
> 93 million oban jobs
haha no way!! whaaaaaaat????In conversation permalink -
Embed this notice
Christmas Sun (sun@shitposter.world)'s status on Friday, 06-Sep-2024 22:27:27 JST Christmas Sun @graf @feld I have under 2,000 oban job rows In conversation permalink -
Embed this notice
anime graf mays ?️? (graf@poa.st)'s status on Friday, 06-Sep-2024 22:27:28 JST anime graf mays ?️? @feld @zop @sun it filled syslog (130+ GB) and pleroma locked up out of disk space
pastebin.com/rQjtzF9bIn conversation permalink Attachments
-
Embed this notice
feld (feld@friedcheese.us)'s status on Friday, 06-Sep-2024 22:28:08 JST feld @graf @zop @sun how is that even possible? How do they successfully deliver that may messages to Pleroma's inbox in such a short period of time? In conversation permalink -
Embed this notice
anime graf mays ?️? (graf@poa.st)'s status on Friday, 06-Sep-2024 22:29:57 JST anime graf mays ?️? @feld @zop @sun I wasn't paying attention to that instance for maybe a day or two, but pleroma seems to be trying to reinsert several times per second per user and per post In conversation permalink -
Embed this notice
feld (feld@friedcheese.us)'s status on Friday, 06-Sep-2024 22:29:57 JST feld @graf @zop @sun we need a new release with some of my fixes in it to discard some of these jobs after certain failures 😭 In conversation permalink -
Embed this notice
feld (feld@friedcheese.us)'s status on Friday, 06-Sep-2024 22:33:54 JST feld @graf @zop @sun since it's only RemoteFetcher it's perfectly harmless to delete everything in the table for that Worker, but make sure you stop Pleroma when you do it.
Time to add more safeguards against this.In conversation permalink -
Embed this notice
anime graf mays ?️? (graf@poa.st)'s status on Friday, 06-Sep-2024 22:33:55 JST anime graf mays ?️? @feld @zop @sun here's a cross section, it's literally the same handful of posts by a handful of users lmfao In conversation permalink Attachments
-
Embed this notice
feld (feld@friedcheese.us)'s status on Friday, 06-Sep-2024 22:48:03 JST feld @graf @zop @skylar @sun
here's a fix that would have at least stopped the duplicates from being accumulated into the queue
https://git.pleroma.social/pleroma/pleroma/-/merge_requests/4253In conversation permalink Attachments
-
Embed this notice
anime graf mays ?️? (graf@poa.st)'s status on Friday, 06-Sep-2024 22:48:04 JST anime graf mays ?️? @skylar @zop @feld @sun i imagine this is going to break a lot of single user instances if they come across these posts tbh In conversation permalink -
Embed this notice
skylar :confederateflag:??? :z: (skylar@misskey.yandere.love)'s status on Friday, 06-Sep-2024 22:48:05 JST skylar :confederateflag:??? :z: @graf @zop @feld @sun my instance was doing the same thing the other day
misskey machine broke?In conversation permalink -
Embed this notice
feld (feld@friedcheese.us)'s status on Saturday, 07-Sep-2024 00:08:15 JST feld @zop @sun dumping RemoteFetcher is always safe, it's just a side effect of processing a thread or fetching users encountered in new posts etc. If you kill it and load the target thread in the future it should generate the required jobs to reconstruct it backfill like it was trying to originally In conversation permalink -
Embed this notice
ZOP (zop@7td.org)'s status on Saturday, 07-Sep-2024 00:08:16 JST ZOP @feld @A3NmXoRtC3rBTN23nM.graf@poa.st @sun Yeah I had...heck 14M, 15M RemoteFetcherWorker's queued. It...looked more like each job created a new job and the original job stayed, but, idk, I didn't have any time to investigate. Couldn't really human read the logs because they were so high volume (milliseconds between entries) so had to turn logging totally off. Didn't know if dumping those jobs was "safe" or not. In conversation permalink -
Embed this notice
PalePimp (palepimp@poa.st)'s status on Saturday, 07-Sep-2024 04:27:33 JST PalePimp @graf @feld @zop @sun Learned the hard way years ago that /home and /var/log go into their own separate volumes on production boxes.
/var/log /dev/mapper/vg_system-lv_logIn conversation permalink Christmas Sun likes this.
-
Embed this notice