-
EyePulp has quit
-
EyePulp joined the channel
-
EyePulp has quit
-
EyePulp joined the channel
-
EyePulp has quit
-
stezzle2 has quit
-
stezzle joined the channel
-
busybox42 has quit
-
stezzle2 joined the channel
-
stezzle has quit
-
stezzle joined the channel
-
stezzle2 has quit
-
stezzle2 joined the channel
-
stezzle has quit
-
stezzle joined the channel
-
stezzle2 has quit
-
stezzle has quit
-
stezzle joined the channel
-
EyePulp joined the channel
-
stezzle2 joined the channel
-
stezzle has quit
-
busybox42 joined the channel
-
brisby has quit
-
stezzle2 has quit
-
stezzle joined the channel
-
stezzle has quit
-
stezzle joined the channel
-
stezzle2 joined the channel
-
stezzle has quit
-
SynchroM joined the channel
-
knutix has quit
-
knutix joined the channel
-
knutix has quit
-
Guest90 joined the channel
-
Guest90 has quit
-
EyePulp joined the channel
-
EyePulp has quit
-
EyePulp joined the channel
-
EyePulp joined the channel
-
EyePulp has quit
-
EyePulp joined the channel
-
baudehlo
_smf_ I got that grafana stuff working. It's a bit of a pain to setup but at least it's there now.
-
Do you just instrument everything with lynx?
-
or something else?
-
_smf_
baudehlo: I just used lynx
-
baudehlo
I can't figure out how to query diskio in grafana. Did you ever do that?
-
_smf_
I'm not monitoring diskio here, but I can't see why you couldn't?
-
baudehlo
It's a standard feature of telegraf, but I can't figure out what the actual query should be.
-
I think I found it.
-
SELECT non_negative_derivative("read_bytes", 10s) FROM "diskio" WHERE $timeFilter
-
knutix joined the channel
-
baudehlo has quit
-
baudehlo joined the channel
-
notkoos_ joined the channel
-
notkoos has quit
-
knutix_ joined the channel
-
knutix has quit
-
knutix_ has quit
-
knutix joined the channel
-
first1
I am supposed to receive e-mail from this domain: telesol.com.gh
-
but it gets rejected with this error by Haraka: MX without A/AAAA records
-
baudehlo
Sounds like your dns is busted.
-
install a local resolver and use that.
-
first1
that's what I am doing
-
local bind cache
-
I am querying against it and it returns results just fine
-
baudehlo
$ host -t mx telesol.com.gh
-
telesol.com.gh mail is handled by 10 mx.telesol.com.gh.
-
$ host mx.telesol.com.gh
-
mx.telesol.com.gh has address 197.159.130.213
-
did you update your resolv.conf?
-
first1
oh yes, it's running the same without issues
-
ah wait
-
now it's failing
-
mx.telesol.com.gh has address 197.159.130.213
-
Host mx.telesol.com.gh not found: 2(SERVFAIL)
-
I wonder what that second part is
-
I need to run tcpdump. the flow of dns queries will kill me
-
12:57:35.068920 IP 192.168.1.139.50535 > 192.168.1.24.domain: 15278+ AAAA? mx.telesol.com.gh. (35)
-
it tries to do an AAAA lookup against that name which gets the SERVFAIL
-
the question is, why does Haraka attempt an AAAA lookup
-
baudehlo
because we support IPv6
-
first1
so AAAA lookup on 8.8.8.8 returns no answer for mx.telesol.com.gh
-
my cache returns SERVFAIL
-
I think there's something wrong with that domain
-
-
origin = dns, that's got to be wrong
-
and even so, shouldn't Haraka return a positive result if at least one of the lookups was good ? In this case, the A lookup
-
first1 is now known as last1
-
last1
looking at the code, the problem appears to be in net_utils
-
in the get_ips_by_host() function
-
it returns the error for both records. I'll try and fix
-
hmm, or not. something is amiss here
-
why would the host command work but nslookup/haraka dns resolution fail ?
-
gah, restarting my bind cache solved the issue but this makes no sense whatsoever
-
baudehlo
Some of the node DNS functions use getaddrbyname and some use internal dns resolver library.
-
That might explain the difference.
-
last1
I restarted bind cache, both host & nslookup return an IP for mx.telesol.com.gh
-
but now Haraka still complains: 450 [BE892A@haraka4] Temp. resolver error (ESERVFAIL)
-
oh yeah, because now my cache returns SERVFAIL for the mx records of telesol.com.gh
-
what do you use as local resolver ?
-
ok, it wasn't a Haraka issue..rather a weird Ghanian issue
-
in fact, it's a Google Public DNS is evil problem
-
baudehlo
you should never use 8.8.8.8 on a mail server
-
last1
I'm not, I use a local resolver
-
but Google is resolving that name and that serves as proof to the remote end that all is good
-
when in fact their bind config is broken
-
bind -> dns
-
SynchroM has quit