Hi mayhem, lucifer. If we can confirm that the tagger in MB android does a better job at identifying now, I suggest we write a reply to the concerning reviews on play store, asking them to reconsider the functionality and update.
outsidecontext, skelly37: we should have a small meeting for next version's plans, along GSoC, what do you think?
atj
zas: we can either remove the monitoring the "failed" disk on trille, or we can remove the "-H" flag, which should stop it notifying about the failed health status
skelly37
zas: fine for me
zas
atj: check /etc/smartmontools/run.d/10mail on trille
(btw, this is the kind of change I would add to ansible)
atj
zas: yes, I saw that, but that's treating the symptom not the cause
we can implement something like that if you really want, but it's a bit of a hack
zas
we can't prevent the cause, the disk actually reports an health failure, it is just that this is due to a buggy firmware (it seems), we can have other valid health failures, so we don't want to silent them all
outsidecontext
zas, skelly37: yes, definitely. I'm a bit low on time this week, though, will be away Thu till Sun
atj
that's what I'm saying, we should disable SMART health status *only* for that drive
"SMART overall-health self-assessment test result: FAILED!"
skelly37
zas, outsidecontext: What about today's evening or tomorrow morning tho?
zas
atj: if we disable SMART health status for this drive, and one is emitted due to another issue, we'll not get the warning at all, the one which has to be ignored is very specific. Since smartmontools doesn't have a way to ignore this specific error, either we disable all, or none. I chose none, and filter out message before it is sent to avoid spamming our mail box.
skelly37: I'm fine with today's evening, or next Tuesday
skelly37
zas: next week i'm fine with wednesday or later, tuesday can't do.
zas
outsidecontext: what about today's evening?
outsidecontext
would work for me today, tomorrow not so much
zas
17 UTC?
skelly37
okay for me
zas
atj: we have to add 5 more servers to ansible, the ones we call Jackson5, those are on 20.04, but they are using a vlan, instead of a physical private network
those are michael, tito, jermaine, jackie, marlon
check their netplan configs (vlans)
outsidecontext
zas, skelly37: 17 UTC is good
zas
atj: I also noted there's a /etc/apt/sources.list.d/adoptopenjdk.list
what's the reasoning being using adopt instead of openjdk?
ah, makes it easier to switch versions?
lucifer
both are same just different packagers. but the official openjdk binaries stop receiving updates after 6 months.
also yes easier to update with adopt.
atj
well, openjdk is bundled with ubuntu so I suspect it receives updates
lucifer
ah ok, i was referring to the official tarballs of openjdk above.
didn't use the openjdk packages in ubuntu because they have a bad rep for breaking stuff.
atj
fair enough
do you really need to run all these java apps as root? :)
lucifer
nope. i didn't know how to setup it as non-root at the time of setting it up initially. i guess it'd be the same but was doing it for first time at that point and just wanted to make it work. 😓
its a pending task to move this to non-root, a better install location (somewhere else than directly under /), add a startup script (maybe systemd) to control the processes.
zas, atj: regarding the jackson5 cluster, some entries need to be added to etc/hosts on each node. the repository as you found and also a ssh key (not sure if this concerns the ansible install). also, please let me know when you do this, because currently the relevant processes don't start automatically on restart.
atj
no worries, I understand the whole "let's just make it work" process
we have /etc/hosts management in ansible now, so we can sort that part
lucifer
ah great!
atj
zas: what other hetzner VMs do we have, aside from these?
zas
7 for SOLR cloud, and few more for various services (jira, livegrep, etc...)
so that's 5 + 7 + ~5
agatzk has quit
agatzk joined the channel
atj
do they have any commonality in the network configuration?
zas
not really, Jackson5 is using a private network using hetzner vlan, solr cloud VMs are using only one IP
solr cloud should be rebuilt on new VMs (new OS + new SOLR°
but jackson5 can be migrated to ansible right now I think. Basically they use a vlan + docker, firewall is ufw, which needs to be migrated to shorewall, but the current config is very simple
atj
I'm creating a new branch to test an approach
alastairp
morning
mayhem
moin!
do you know if monkey is coming in today? if so, we could have an oauth chat
alastairp
monkey: will we be seeing you at the office today? got some things to run past you
great question,
mayhem
lol
monkey
Wasn't planning on it, but can do
Why not. Are we all having lunch as well?
mayhem
not critical from my side, but was just wondering. could do later in the week with less threat of rain
I'm up for that!
monkey
Almost-rainy day is perfect for talking sho^
shop
mayhem
not so great for cycling to office.
then again, you have the office express to rely on.
(L1)
monkey
Choo choo
The great thing about the metro is that the book you're reading won't get wet.
alastairp
I have lunch with me today!
mayhem
alastairp: what if you get stood up?
alastairp
but I'll probably come in tomorrow as well, if it's not raining we could do lunch and talking then
mayhem
I'm not available for lunch tomorrow. tomorrow is a rather choppy day for me.
alastairp
thurs/fri isn't good for me for office in general
in terms of getting together to chat about oauth
mayhem
I guess it would be good for you to come in then monkey.