libre #smartos

16 Dec 2022

SmartOS -- A Modern, Open Source, Datacenter Operating System | smartos.org | github.com/TritonDataCenter/smartos-live | github.com/TritonDataCenter/triton | Security -- TritonDataCenter.com/security-issues | Channel logs - log.omnios.org/smartos

• 00:14
  Smithx10
  lol.... this is fun..... github.com/influxdata/go-syslog/blob/develop/rfc5424/machine.go#L129
• 00:33
  jbk
  i can only hope that's generated code from some other input (ala yacc/lex)
• 02:32
  Smithx10
  jbk: I think its generated by something called "ragel"
• 03:34
  rennj
  you mean using computers to autom8 repetitive tasks...i dont believe it!
• 03:34
  rennj
  614 case statements
• 03:36
  rennj
  testof big deal...
• 03:50
  Smithx10
  rennj: lol
• 03:51
  Smithx10
  I'm confused at why when we move messages into /var/log/messages those fields are cut out lol
• 03:52
  Smithx10
  or maybe illumos no follow the rfc
• 03:52
  Smithx10
  Didn't think I'd be playing with ragel today :)
• 04:03
  Smithx10
  : ixgbe3: Could not allocate rbd dma memory: ffffffff ixgbe3: Failed to allocate DMA resource..... this doesn't seem good
• 04:51
  jbk
  if you have VMs you're not using, try shutting them down
• 04:52
  jbk
  hmm.. i actually have ixgbe nics...
• 04:52
  jbk
  so if no one else gets to it in the next few months, i'll probably try to fix that after i've moved and have a 10gb switch to use w/ it
• 04:57
  rennj
  the switch being the important part
• 04:57
  rennj
  1 ended is useless
• 04:59
  rennj
  AGI and starving artist!
• 05:01
  rennj
  so vmware HA/vmotion between dc's takes minutes to failover...lame
• 05:01
  rennj
  not seconds but minutes
• 05:07
  rennj
  the computer industry/rent seeking has left me jaded
• 05:10
  rennj
  its all SaaS
• 05:13
  arekinath
  my personal fav is when vmware VMs just sometimes... stop being able to vmotion at all
• 05:13
  arekinath
  I have like 3 of them where vmotion just errors out with weird error codes that google doesn't know
• 05:15
  arekinath
  there's a reason people say you should never rely on hypervisor-layer stuff for actual HA in a failure scenario (vs. planned outage failover), heh
• 05:15
  jbk
  i love when if for whatever reason your NFS datastore times out.. you have to reboot the ESX host
• 05:15
  rennj
  you are only as good as your backups
• 05:15
  jbk
  you can't just reset the VMs or anything once the NFS mount is back
• 05:16
  rennj
  52weeks level 0, 5,4,3,2,1-->0... weekly per server...you need 57tapes
• 05:16
  rennj
  just to do yearly backups before the ransom ware encrypts your data
• 05:17
  rennj
  52 tapes for just level 0, rotate the other 5 tapes till failure
• 05:18
  rennj
  my storagetek/adic is screwed in end
• 05:19
  rennj
  robot library is not going to fix the data issue...
• 05:20
  rennj
  bitcoin worse then mining actual gold..for env
• 05:26
  rennj
  yeah well lameness rules at this point
• 05:27
  rennj
  Roko's basilisk suffer...
• 05:28
  rennj
  its all dependency hell
• 05:31
  rennj
  draw me mona lisa in DaVinci style...
• 05:35
  rennj
  and dell/emc will sell vmware to broadcom for what $61 billion
• 13:47
  neuroserve
  dl.gi.de/bitstream/handle/20.500.12116/38504/Paper01.pdf <- Operating System Name Options exhausted?
• 14:18
  jlevon
  you think they'd google
• 14:32
  andyf
  There's an OmniOS too, somewhere... I don't think they bother checking!
• 15:15
  Smithx10
  Would only having 14gb free of memory on a machine make an interface flap?
• 15:16
  Smithx10
  or I guess running low on memory could do this?
• 15:19
  Smithx10
  0 0 687 753157100 4724428 2211 34702 1 0 0 0 0 150 185 185 149 139862 153507 191941 2 17 81 I have a machine with 4gb of memory free and the interfaces is up/downing like crazy
• 15:35
  Smithx10
  I freed up some memory and got it to 24GB in vmstat, the up/down is still going strong, not sure if its related at all
• 15:48
  danmcd
  I'm imagining fragmentation for some reason.
• 15:49
  jbk
  I don't think it'd make it flap (link up/down).. IIRC that's normally just an indication from the HW
• 15:50
  jbk
  the big issue with fragmentation is usually with jumbo frames -- since that means trying to find potentially lots of chunks of physically contiguous memory
• 15:50
  danmcd
  Oh wait this is ixgbe not i40e as well. (i40e is a pig whenever you bring up a new vnic.)
• 15:52
  danmcd
  Okay, more likely a HW issue then.
• 15:56
  Smithx10
  We do have the underlay using 9k but im not sure if that wuld cause it.... going to have DC folks go replace cables and audit more *
• 16:11
  Smithx10
  interesting
• 16:11
  Smithx10
  I removed the interface from the aggregation and the up/downs stopped
• 16:13
  jbk
  do you see anything on the switch side?
• 16:14
  Smithx10
  I dont have access :(
• 16:14
  Smithx10
  this is the problem
• 16:27
  danmcd
  Oh yeah, you have one of those "traditional" network vs. machines splits. :barf:
• 16:30
  jbk
  and of course the network is never the problem.. the network is flawless... perfection manifest... all who gaze upon it despair...
• 16:37
  jbk
  also.. thank god for the various unix shell utilities