Closed jvanbaarsen closed 4 years ago
Any updates on this?
@kjeldahl could you paste the contents of /proc/diskstats
here? Feel free to anonimize if needed, but an example of what it looks like on your system would be very useful.
Nevermind, just received this example from a colleague:
202 0 xvda 11514 271 1204763 14486 130975 18219 2798097 178789 0 42381 112571 0 0 0 0 0 0
202 1 xvda1 169 34 8894 115 10 6 44 7 0 99 51 0 0 0 0 0 0
202 2 xvda2 258 216 11969 238 1 0 1 0 0 98 135 0 0 0 0 0 0
202 3 xvda3 26 0 208 13 0 0 0 0 0 13 5 0 0 0 0 0 0
202 4 xvda4 11004 21 1181180 14094 130964 18213 2798052 178781 0 42317 112377 0 0 0 0 0 0
202 80 xvdf 25663 55 810823 29159 33077 21044 1430229 37739 0 42594 33450 0 0 0 0 0 0
202 81 xvdf1 25621 55 808479 29142 33077 21044 1430229 37739 0 42583 33447 0 0 0 0 0 0
I investigated this a bit last night, and it would seem that kernel 5.5+ returns 20 segments (we look for 14 or 18 currently). I'm working on a fix for this.
cc/ @thijsc
@thijsc Any news on when this is updated in appsignal-ruby?
@kjeldahl it should be included in the Ruby gem since the 2.11.1 release.
A customer reported that they had issues with running the agent on Alpine Linux. They get the error:
[2020-05-01T10:17:00 (agent) #18][Error] Error getting disk IO stats: Incorrect number of segments
.The error is thrown by this code: https://github.com/appsignal/probes-rs/blob/5ddbd5eeb88f0daa72a5d7f08e39f0953a0dae3d/src/disk_stats.rs#L110.
Intercom conversation is here: https://app.intercom.com/a/apps/yzor8gyw/inbox/inbox/2788926/conversations/21831342893