[00:03:33] shdubsh: hi, if you're around, do you know anything on top of your head about issues like this? T402247 [00:03:34] T402247: rsyslog is segfaulting non-stop on ms-be1071 - https://phabricator.wikimedia.org/T402247 [00:10:17] Amir1: I think it's because the file descriptor for '/var/spool/rsyslog/centrallog1002.eqiad.wmnet:6514.00006358' is open but the file no longer exists. [00:11:42] denisse: thanks! this really shouldn't trigger a segfault but meh. I should force close the fd? I admit I've never done stuff like this before [00:12:16] Amir1: I'm investigating, in theory restarting the unit must suffice if that was the case, but it didn't work. [00:12:58] thanks! [00:20:20] Amir1: Running it debug mode: https://www.irccloud.com/pastebin/jqQpq8r2/ [00:20:34] I'm not sure if it could be because of corrupt data. [00:21:21] honestly at this point, starting afresh is fine, a lot of logs have been lost anyway [00:21:42] (if the log file is corrupted) [00:22:31] Amir1: Okay, let me try to get it working. [00:28:53] Amir1: I wonder if the disk itself could be failing, because I see these logs: [37968216.180582] blk_update_request: I/O error, dev sdg, sector 348391680 op 0x0:(READ) flags 0x1000 phys_seg 4 prio class 0 [00:29:30] could be [00:30:15] It's 2:30am here, my brain is fried. Would you mind posting in the ticket, we can take a look tomorrow [00:30:29] Sure, rest well.