Long time span with no update in log while gc is running...

[mhert]

Is it normal that there are huge time spans where no update seems to happen in logfile (last three lines):

2024-10-25T08:21:40+02:00: starting garbage collection on store remote_backupstorage_nas07
2024-10-25T08:21:40+02:00: Start GC phase1 (mark used chunks)
2024-10-25T08:23:12+02:00: marked 1% (10 of 921 index files)
2024-10-25T08:23:38+02:00: marked 2% (19 of 921 index files)
2024-10-25T08:24:04+02:00: marked 3% (28 of 921 index files)
2024-10-25T08:25:34+02:00: marked 4% (37 of 921 index files)
2024-10-25T08:25:59+02:00: marked 5% (47 of 921 index files)
2024-10-25T08:26:27+02:00: marked 6% (56 of 921 index files)
2024-10-25T08:27:58+02:00: marked 7% (65 of 921 index files)
2024-10-25T08:28:10+02:00: marked 8% (74 of 921 index files)
2024-10-25T08:28:21+02:00: marked 9% (83 of 921 index files)
2024-10-25T08:31:50+02:00: marked 10% (93 of 921 index files)
2024-10-25T08:33:50+02:00: marked 11% (102 of 921 index files)
2024-10-25T08:38:54+02:00: marked 12% (111 of 921 index files)
2024-10-25T08:44:42+02:00: marked 13% (120 of 921 index files)
2024-10-25T08:48:19+02:00: marked 14% (129 of 921 index files)
2024-10-25T08:48:36+02:00: marked 15% (139 of 921 index files)
2024-10-25T08:48:53+02:00: marked 16% (148 of 921 index files)
2024-10-25T08:55:10+02:00: marked 17% (157 of 921 index files)
2024-10-25T09:01:30+02:00: marked 18% (166 of 921 index files)
2024-10-25T09:07:58+02:00: marked 19% (175 of 921 index files)
2024-10-25T09:18:43+02:00: marked 20% (185 of 921 index files)
2024-10-25T09:25:57+02:00: marked 21% (194 of 921 index files)
2024-10-25T09:32:14+02:00: marked 22% (203 of 921 index files)
2024-10-25T09:40:12+02:00: marked 23% (212 of 921 index files)
2024-10-25T09:45:50+02:00: marked 24% (222 of 921 index files)
2024-10-25T09:50:32+02:00: marked 25% (231 of 921 index files)
2024-10-25T09:54:57+02:00: marked 26% (240 of 921 index files)
2024-10-25T09:55:20+02:00: marked 27% (249 of 921 index files)
2024-10-25T09:55:31+02:00: marked 28% (258 of 921 index files)
2024-10-25T09:55:43+02:00: marked 29% (268 of 921 index files)
2024-10-25T09:56:06+02:00: marked 30% (277 of 921 index files)
2024-10-25T09:56:17+02:00: marked 31% (286 of 921 index files)
2024-10-25T09:56:27+02:00: marked 32% (295 of 921 index files)
2024-10-25T09:57:14+02:00: marked 33% (304 of 921 index files)
2024-10-25T09:58:41+02:00: marked 34% (314 of 921 index files)
2024-10-25T10:00:00+02:00: marked 35% (323 of 921 index files)
2024-10-25T10:01:32+02:00: marked 36% (332 of 921 index files)
2024-10-25T10:01:45+02:00: marked 37% (341 of 921 index files)
2024-10-25T10:02:00+02:00: marked 38% (350 of 921 index files)
2024-10-25T10:02:16+02:00: marked 39% (360 of 921 index files)
2024-10-25T10:03:04+02:00: marked 40% (369 of 921 index files)
2024-10-25T10:03:18+02:00: marked 41% (378 of 921 index files)
2024-10-25T10:03:33+02:00: marked 42% (387 of 921 index files)
2024-10-25T10:04:10+02:00: marked 43% (397 of 921 index files)
2024-10-25T10:04:25+02:00: marked 44% (406 of 921 index files)
2024-10-25T10:04:46+02:00: marked 45% (415 of 921 index files)
2024-10-25T10:05:18+02:00: marked 46% (424 of 921 index files)
2024-10-25T10:05:28+02:00: marked 47% (433 of 921 index files)
2024-10-25T10:05:37+02:00: marked 48% (443 of 921 index files)
2024-10-25T10:58:18+02:00: marked 49% (452 of 921 index files)
2024-10-25T15:22:48+02:00: marked 50% (461 of 921 index files)

Should the time per percent not be almost equal? No other things are running on pbs / the storage at the same time...

 

[dcsapak]

well this depends a bit on the number of chunks in the index file. since the garbage collection does not know in advance how many chunks there are, the percentage is in regards to the snapshot indices, but those can contain a different amount of chunks of course. so if you have some big backups those will naturally take longer to go through...

EDIT: Also of course it depends on the speed of the storage, so if at some point there is some additional load on it, during those the GC will take longer

 

[mhert]

thank you for clarification.