Refreshrefreshrefreshrefresh
-
@candlejack1 said in Refreshrefreshrefreshrefresh:
@ben_lubar if you need help to write a 5 line bash script that restart the dockers when the cpu is at 100% for more than 30s just say the word
-
@FrostCat said in Refreshrefreshrefreshrefresh:
Has this actually been reported--I'm not suggesting you, Blakey, do it--to the nodebb devs?
Sure:
-
@Yamikuronue said in Refreshrefreshrefreshrefresh:
@FrostCat said in Refreshrefreshrefreshrefresh:
Has this actually been reported--I'm not suggesting you, Blakey, do it--to the nodebb devs?
Sure:
I'd reply there but then I'd have to remember my password. Instead of a toaster, I'd rather see a text message, maybe next to the submit button--and/or even the submit button changing color significantly or something (that is, rather than just the faded blue it gets now when you click it, actually becoming grey or red or something. (I know you mentioned "greying out" in your report but I want to clarify what I mean.)
-
@boomzilla said in Refreshrefreshrefreshrefresh:
I could possibly see that happening if you were in the middle of a request right at the point where the containers did the switch over. I don't know enough about how those sorts of things work on that level to say one way or the other.
This exact process happens to me every day. Pages stop loading, a refresh shows a 502 for a minute or so, then it's back with an update post. And again, all I'm asking for is a scheduled time for it. Even the refresh is annoying multiple times a day.
-
@Magus said in Refreshrefreshrefreshrefresh:
Pages stop loading, a refresh shows a 502 for a minute or so, then it's back with an update post.
oddly, the root cause of this is: the toaster lies to you.
no, bear with me for a moment.
the toaster lies to you, it assumes that the only reason the instance was restarted was because updates were applied. so when we restart the server to clear the cooties it tells you that updates were applied. It doesn't actually bother checking to see if updates were applied it just says "oh, server went away, then came back. must have been updates."
see? it lies to you.
it desn't help that @ben_lubar applies updates any time he feels like it (seriously ben, can we keep it down to at most once a day, preferably once a week? thanks) but the root cause of your annoyance seems to be the lieing toaster
-
@accalia said in Refreshrefreshrefreshrefresh:
the toaster lies to you, it assumes that the only reason the instance was restarted was because updates were applied. so when we restart the server to clear the cooties it tells you that updates were applied. It doesn't actually bother checking to see if updates were applied it just says "oh, server went away, then came back. must have been updates."
Oh seriously!?
Jeez.
-
@accalia Um, actually, I believe it uses a "cache buster" UUID-like thing to identify that a refresh is required.
-
@error said in Refreshrefreshrefreshrefresh:
@accalia Um, actually, I believe it uses a "cache buster" UUID-like thing to identify that a refresh is required.
a cachebuster uuid that changes when the instance restarts, yes.
it'll not trigger for a websocket collapse, but it will trigger for the server disappearing, then reappearing
@blakeyrat said in Refreshrefreshrefreshrefresh:
Oh seriously!?
ya srsly
-
@accalia said in Refreshrefreshrefreshrefresh:
@error said in Refreshrefreshrefreshrefresh:
@accalia Um, actually, I believe it uses a "cache buster" UUID-like thing to identify that a refresh is required.
a cachebuster uuid that changes when the instance restarts, yes.
That's what I was saying (going for that ).
Of course, it would be better if the dist build just spit out the md5 of the current changeset and they used that as the cache buster. They could at least vary the message appropriately.
The server has itself.
-
@accalia I said "update post" not notification or toaster. So good job. I just want these cases to happen once a day at a scheduled time. How many times do I have to say it?
It's not complicated. It's not even hard.
-
@Magus said in Refreshrefreshrefreshrefresh:
. How many times do I have to say it?
how many times do we all have to say it until @ben_lubar realizes that he is the one that's causing all the salt by updating whenever he feels like it.
this is a production server there should be a SCHEDULED update window. it should be ON THE WEEKEND and it should have a ROLLBACK PLAN in case of issues.
this is how things should work, and it's more or less how they worked in discourse towards the end there when @PJH was doing the update (well except for the rollback plan, because no one seems to understand what that's for nor why it should be desireable)
this is not how @ben_lubar seems to work, and that's a shame, but it is not my shame, nor is it your shame. no the shame is @ben_lubar's and it is a shame he will carry with him e'en unto Gre'thor unless he repents and steers us unto the path of the scheduled maintenance window. he must atone for his sins of random updates or ne'er will he see Sto-vo-kor.
Because this is not my dishonour, and because my pleas have fallen upon deaf ears, i have resigned myself to the current state of affairs. Either Ben will learn, or he will not, but i will not waste my effort on trying to teach him who won't listen to his constituants, or if he is listening one who would listen to our pleas and ignore them.
-
@Magus said in Refreshrefreshrefreshrefresh:
This exact process happens to me every day. Pages stop loading, a refresh shows a 502 for a minute or so, then it's back with an update post. And again, all I'm asking for is a scheduled time for it.
And again, let me know when you've figured out how to schedule server cooties storms and we'll look into that.
-
@boomzilla said in Refreshrefreshrefreshrefresh:
And again, let me know when you've figured out how to schedule server cooties storms and we'll look into that.
I have yet to ask for that. I only ask that the updates be scheduled. That's all I ever have asked for, despite your expert misreading skills.
-
@Magus said in Refreshrefreshrefreshrefresh:
How many times do I have to say it?
No one knows why you keep saying it. Really, it's a mystery.
-
@Magus said in Refreshrefreshrefreshrefresh:
I have yet to ask for that.
Except that the situation you keep describing is that.
-
By the way, the spinner lives here:
That's hover-text, it's usually just the spinner
-
@Magus said in Refreshrefreshrefreshrefresh:
It's not even hard.
Just relax and ease into it. Don't put too much pressure on yourself.
-
@HardwareGeek said in Refreshrefreshrefreshrefresh:
immediately followed by
So much for "zero downtime."
To be fair, I'm kinda glad that one was applied promptly...
And not even because I reported that bug first.
-
@FrostCat I just made a RFC about it in the NodeBB community forum. Hopefully that gets discussion going
-
@boomzilla said in Refreshrefreshrefreshrefresh:
Except that the situation you keep describing is that.
I describe that after an upgrade happens, there are cooties. This is a fact. To isolate it, and because it's the sane thing to do, I want updates scheduled.
Whatever else you're adding on is your problem.
-
@Magus Not exactly.
More accurately, the "updates" are caused by the cootie storms. The "the server has been updated" notifications are sent whenever the server reboots (the toaster should probably be changed to reflect that?).
So usually, it's not an "update", but just a reboot caused by a pegged core.
-
@sloosecannon said in Refreshrefreshrefreshrefresh:
@Magus Not exactly.
More accurately, the "updates" are caused by the cootie storms. The "the server has been updated" notifications are sent whenever the server reboots (the toaster should probably be changed to reflect that?).I specifically am referring to when there is a cootie storm and an update post on the updates thread.
As I've said.
Over.
And over.
And over.
-
@Magus said in Refreshrefreshrefreshrefresh:
I describe that after an upgrade happens, there are cooties.
Nope. The changeover on an update is generally pretty smooth. You should get a toaster, but no cooties. I guess it's possible that the "old" container was experiencing cooties when an update was happening, but that's a separate issue and probably would have happened regardless (though I can't rule out that extra load on the server from spinning up a new container might not contribute or precipitate a storm).
-
@Magus
I personally haven't seen a cootiestorm after the updates, but I may just not have seen it.The updates are pretty infrequent though, and not the biggest cause for the downtime.
However, it's probably not a bad idea for @ben_lubar to refrain from causing more downtime during business hours unless we really have to...
-
@boomzilla said in Refreshrefreshrefreshrefresh:
Nope. The changeover on an update is generally pretty smooth.
Prove it. Schedule it. Gather metrics. I flat out don't believe you, because all our updates are completely randomly timed, whenever ben feels like it, and I have observed the opposite.
-
So what happens exactly in the time leading up to a post getting added to https://what.thedailywtf.com/topic/19454/nodebb-updates/1 ?
-
@aliceif I get a spinner and 502s, and apparently none of the admins touch the forum in any way.
-
@Magus said in Refreshrefreshrefreshrefresh:
@boomzilla said in Refreshrefreshrefreshrefresh:
Nope. The changeover on an update is generally pretty smooth.
Prove it. Schedule it. Gather metrics. I flat out don't believe you, because all our updates are completely randomly timed, whenever ben feels like it, and I have observed the opposite.
Paging @accalia - what does servercooties say about the update at 11:09:22 EST today?
-
Seriously, just schedule the updates. Why do you have a problem with this? Because you'd prove me right? If you're right, we still get an improvement!
-
@Magus said in Refreshrefreshrefreshrefresh:
Prove it. Schedule it. Gather metrics. I flat out don't believe you, because all our updates are completely randomly timed, whenever ben feels like it, and I have observed the opposite.
LOL, fuck off. This is your theory. You should be convincing me. Get the servercooties archive and correlate the updates with the cooties storm.
I get that you have less visibility than I do into what's happening when on this site and how it could confuse you about what's what.
-
@boomzilla said in Refreshrefreshrefreshrefresh:
LOL, fuck off. This is your theory. You should be convincing me. Get the servercooties archive and correlate the updates with the cooties storm.
I get that you have less visibility than I do into what's happening when on this site and how it could confuse you about what's what.Look, Jeff, you have user reports. Lots of them. Complaints. What are you going to do with them?
- Ignore
- Deny
- Tell the user they're looking at it wrong
Three for three! Well done, Jeff!
-
@boomzilla said in Refreshrefreshrefreshrefresh:
I get that you have less visibility than I do into what's happening when on this site and how it could confuse you about what's what.
Separating it so this doesn't happen is my feature request. Even if it doesn't solve the cooties. Do it for just this reason and I'm satisfied.
-
@Magus said in Refreshrefreshrefreshrefresh:
I get a spinner and 502s,
Yesterday the server was keeping it fresh by mixing 504s in with the 502s.
-
@Magus said in Refreshrefreshrefreshrefresh:
Look, Jeff, you have user reports. Lots of them. Complaints. What are you going to do with them?
Ignore
Deny
Tell the user they're looking at it wrongThree for three! Well done, Jeff!
Good lord you're obstinate when you're wrong.
@Magus said in Refreshrefreshrefreshrefresh:
Separating it so this doesn't happen is my feature request. Even if it doesn't solve the cooties. Do it for just this reason and I'm satisfied.
Separate what? Back off on the whargarbl and read what I've been posting.
-
@boomzilla said in Refreshrefreshrefreshrefresh:
Separate what? Back off on the whargarbl and read what I've been posting.
Most of it seemed to be "THE TOASTER DOESN'T MEAN UPDATES" which had nothing to do with what @magus said.
-
@aliceif said in Refreshrefreshrefreshrefresh:
@boomzilla said in Refreshrefreshrefreshrefresh:
Separate what? Back off on the whargarbl and read what I've been posting.
Most of it seemed to be "THE TOASTER DOESN'T MEAN UPDATES" which had nothing to do with what @magus said.
The toaster means updates since the update where I made the cache buster not change without a docker container change.
-
@boomzilla said in Refreshrefreshrefreshrefresh:
Separate what? Back off on the whargarbl and read what I've been posting.
The. Updates. From. Normal. Activity.
You know, the thing I've been asking for, I don't know, the whole thread long, and that you seem intent on ignoring.
Why not do it? Why is scheduling updates a bad idea? Irrespective of anything else, I want that.
-
"Hey, can we do <sane thing>? I have a feeling that not doing it is causing <problem>, but I'm not really sure."
"ACTUALLY, <problem> happens at random times!"
"Great, but can we <sane thing>?"
"<problem> is based on <blah> and is <bleh's fault!"
"Right, I get it, but <sane thing> will ensure we don't get them confused, and is a good idea anyway..."
"How could you complain about <sane thing>!? Obviously it works exactly as expected and is totally unrelated in every way from <problem>. Don't even say the two in the same sentence. In fact, you aren't even saying things I can read."
"Do. <Sane. Thing.>!"
-
@sloosecannon said in Refreshrefreshrefreshrefresh:
Paging @accalia - what does servercooties say about the update at 11:09:22 EST today?
https://servercooties.io/static/servercooties.sql.zip
have fun.
-
@accalia said in Refreshrefreshrefreshrefresh:
a cachebuster uuid that changes when the instance restarts, yes.
is it possible to use imageid instead of containerid?
-
@candlejack1 said in Refreshrefreshrefreshrefresh:
@accalia said in Refreshrefreshrefreshrefresh:
a cachebuster uuid that changes when the instance restarts, yes.
is it possible to use imageid instead of containerid?
How would that help? They change at exactly the same rate.
-
Since there seems to be a notification of an update at the top today, perhaps my repetition has finally paid off; if so, this thread can be closed or not, but I have no further points to push.
-
@ben_lubar dunno why any of these would change in a restart
-
@accalia said in Refreshrefreshrefreshrefresh:
@sloosecannon said in Refreshrefreshrefreshrefresh:
Paging @accalia - what does servercooties say about the update at 11:09:22 EST today?
https://servercooties.io/static/servercooties.sql.zip
have fun.
ben@australium:~$ mkdir servercooties ben@australium:~$ cd servercooties/ ben@australium:~/servercooties$ ls ben@australium:~/servercooties$ wget https://servercooties.io/static/servercooties.sql.zip --2016-07-01 21:13:18-- https://servercooties.io/static/servercooties.sql.zip Resolving servercooties.io (servercooties.io)... 45.55.146.244, 2604:a880:800:10::76a:f001 Connecting to servercooties.io (servercooties.io)|45.55.146.244|:443... connected. HTTP request sent, awaiting response... 200 OK Length: 69915409 (67M) [application/zip] Saving to: ‘servercooties.sql.zip’ servercooties.sql.z 100%[===================>] 66.68M 1018KB/s in 53s 2016-07-01 21:14:11 (1.26 MB/s) - ‘servercooties.sql.zip’ saved [69915409/69915409] ben@australium:~/servercooties$ unzip servercooties.sql.zip Archive: servercooties.sql.zip inflating: var/www/SockSite/servercooties.sql ben@australium:~/servercooties$ sqlite3 var/www/SockSite/servercooties.sql SQLite version 3.11.0 2016-02-15 17:29:24 Enter ".help" for usage hints. sqlite> .tables checks2 pages sqlite> select count(*) from checks2; Error: database disk image is malformed
-
It's not a real cootiestorm unless it lasts over an hour. The ones that only last a minute are more a cootiesquallline…
-
@ben_lubar Was wondering how you got unicode characters onto the forum.
Then I figured it out...
-
@ben_lubar said in Refreshrefreshrefreshrefresh:
When you say "fun", do you mean ?
....... huh.....
what?
accalia_de_elementia@sockdrawer:~/workspace/servercooties $ wget https://servercooties.io/static/servercooties.sql.zip --2016-07-02 11:36:20-- https://servercooties.io/static/servercooties.sql.zip Resolving servercooties.io (servercooties.io)... 45.55.146.244, 2604:a880:800:10::76a:f001 Connecting to servercooties.io (servercooties.io)|45.55.146.244|:443... connected. HTTP request sent, awaiting response... 200 OK Length: 70060232 (67M) [application/zip] Saving to: ‘servercooties.sql.zip’ 100%[===============================================================================================================================================================================================================================>] 70,060,232 69.4MB/s in 1.0s 2016-07-02 11:36:21 (69.4 MB/s) - ‘servercooties.sql.zip’ saved [70060232/70060232] accalia_de_elementia@sockdrawer:~/workspace/servercooties $ unzip servercooties.sql.zip Archive: servercooties.sql.zip inflating: var/www/SockSite/servercooties.sql accalia_de_elementia@sockdrawer:~/workspace/servercooties $ sqlite3 var/www/SockSite/servercooties.sql SQLite version 3.8.2 2013-12-06 14:53:30 Enter ".help" for instructions Enter SQL statements terminated with a ";" sqlite> select count(*) from checks2; 7074537 sqlite> select * from checks2 order by oid desc limit 10; 378|200|372|1467432014728 377|200|444|1467432009336 1|200|419|1467432003860 378|200|410|1467431998424 377|200|434|1467431992988 1|200|537|1467431987540 378|200|396|1467431981983 377|200|427|1467431976562 1|200|565|1467431971117 378|200|387|1467431965538 sqlite> accalia_de_elementia@sockdrawer:~/workspace/servercooties $
-
Why is everyone posting console logs all the sudden?
-
@Tsaukpaetra said in Refreshrefreshrefreshrefresh:
Why is everyone posting console logs all the sudden?
Easier than pasting screenshots with Firefox. ;)
-
Mobile was actually working well for me yesterday morning (apart from the usual spastic UI dumbfuckery). For a while. Then around noon, the Reply buttons stopped working, and NO amount of refreshing, closing and reopening tabs, nor anything else could get it to work again.