Description of problem: Seems like host/hr_table is broken in 5.1 on linux... hrStorageSize (5), hrStorageUsed (6) are missing. Steps to Reproduce: 1. (size) snmpwalk -v 2c -c foo localhost .1.3.6.1.2.1.25.2.3.1.5 (used) snmpwalk -v 2c -c foo localhost .1.3.6.1.2.1.25.2.3.1.5 Actual results: Prints NOSUCHINSTANCE Expected results: hrStorageSize should print out something similar to: % snmpwalk -O n -v 2c -c foo localhost .1.3.6.1.2.1.25.2.3.1.5 .1.3.6.1.2.1.25.2.3.1.5.1 = INTEGER: 251006 .1.3.6.1.2.1.25.2.3.1.5.2 = INTEGER: 26266935 .1.3.6.1.2.1.25.2.3.1.5.3 = INTEGER: 504011 .1.3.6.1.2.1.25.2.3.1.5.4 = INTEGER: 1512088 .1.3.6.1.2.1.25.2.3.1.5.5 = INTEGER: 504004 .1.3.6.1.2.1.25.2.3.1.5.6 = INTEGER: 29542469 .1.3.6.1.2.1.25.2.3.1.5.101 = INTEGER: 1548036 .1.3.6.1.2.1.25.2.3.1.5.102 = INTEGER: 2048248 and hrStorageUsed % snmpwalk -O n -v 2c -c foo localhost .1.3.6.1.2.1.25.2.3.1.6 .1.3.6.1.2.1.25.2.3.1.6.1 = INTEGER: 53895 .1.3.6.1.2.1.25.2.3.1.6.2 = INTEGER: 17009811 .1.3.6.1.2.1.25.2.3.1.6.3 = INTEGER: 8207 .1.3.6.1.2.1.25.2.3.1.6.4 = INTEGER: 266372 .1.3.6.1.2.1.25.2.3.1.6.5 = INTEGER: 25071 .1.3.6.1.2.1.25.2.3.1.6.6 = INTEGER: 2730738 .1.3.6.1.2.1.25.2.3.1.6.101 = INTEGER: 1537432 .1.3.6.1.2.1.25.2.3.1.6.102 = INTEGER: 199296
Created attachment 96347 [details] Oneliner to enable disk volume sizes/usage again Seems to work for me.
Created attachment 96353 [details] Adds output of memory buffers in use Ok. The real problem here was that if NO_DUMMY_VALUES is in use (it's the default) var_hrstorage() returns NULL instead of *u_char. Somehow this hoses everything in the calling layer above and no information further down the oid row is will be returned. I got bored determining what's calling what and decided to fix this issue another way -- by adding to the missing functionality on linux preventing the return of NULL. :) Works for me, now, really.
Created attachment 96354 [details] Return total and buffers (instead of free memory and buffers) I was figuring buffers would have been allocated from memory free but this wasn't the case and it resulted in funny looking graphs (say there's only 10 M free but 90M buffers you end up with 900% as an end result).. instead we return the total amount of memory and the amount of buffers allocated. Hopefully this will end with values 0 <= x <= 100 when dividing hrStorageUsed with hrStorageSize.
Yeah, that looks good. Have you submitted the patch to Wes or David upstream? Might be a good idea if not. ;-) Thanks again, Read ya, Phil PS: And of course it will be included in the next net-snmp-5.1-3 package. :-)