Bug in FE Update 2 Date Command

A forum for general AmigaOS 4.x support questions that are not platform-specific
User avatar
nbache
Beta Tester
Beta Tester
Posts: 1714
Joined: Mon Dec 20, 2010 7:25 pm
Location: Copenhagen, Denmark
Contact:

Re: Bug in FE Update 2 Date Command

Post by nbache »

Yeah, I'm sorry that the fix isn't released right away, unfortunately it's out of my hands.

As a workaround, you could consider using a more reliable time server.

I have never had this problem with my own ISP's NTP server, and I haven't heard about others having it, except for you.

Best regards,

Niels
dstastny
Posts: 48
Joined: Fri Dec 16, 2016 6:31 am
Location: Atlanta GA

Re: Bug in FE Update 2 Date Command

Post by dstastny »

@Neils,

I am pretty sure its not the time server 1.north-america.pool.ntp.org, I reboot my mesh router every night and think its a time overlap that network is just down. So stumbled into this one :lol:

I think I will gap my scheduled updates to not worry about being correct from 2am - 6am. Hopefully I will be sleeping.

Doug
User avatar
nbache
Beta Tester
Beta Tester
Posts: 1714
Joined: Mon Dec 20, 2010 7:25 pm
Location: Copenhagen, Denmark
Contact:

Re: Bug in FE Update 2 Date Command

Post by nbache »

Ah, I see.

Well, if you know how long it can take in the worst case to reboot your router, maybe you can just schedule your NTP updates outside that. E.g. if you reboot the router at 2:00 am, and it never takes more than 5 minutes, schedule the NTP updates at 15 minutes after the hour.

Best regards,

Niels
MikeSteed
Posts: 11
Joined: Sat Aug 10, 2013 7:07 am

Re: Bug in FE Update 2 Date Command

Post by MikeSteed »

I have never had this problem with my own ISP's NTP server, and I haven't heard about others having it, except for you.
And me! That's how I discovered the bug in the first place. I use a similar NTP server to dstastny -- us.pool.ntp.org -- to set the date and time from Network-Startup every time I boot. It works the large majority of the time, but every now and then (once a month, maybe) it happens to be offline just when I'm booting, and that's when I get the crash. The last time was just a few days ago.

Mike
User avatar
Raziel
Posts: 1170
Joined: Sat Jun 18, 2011 4:00 pm
Location: a dying planet

Re: Bug in FE Update 2 Date Command

Post by Raziel »

Thank you *very* much for this thread and description of the problem.

I was scratching my head for a long time, trying to find out why my system randomly stalls during boot.
I did everything to find out and came close to the date command, but never blamed it.

My problem was, that i had a >NIL; after the command, so it never fired or printed anything.

I also didn't get a crash first, just a *a lot* of characters printed in shell together with *lots* of error beeps...for a *very* long time (before i gave up and broke it).

The second time i tried i got the crash.

Anyone got a temporary workaround?
People are dying.
Entire ecosystems are collapsing.
We are in the beginning of a mass extinction.
And all you can talk about is money and fairytales of eternal economic growth.
How dare you!
– Greta Thunberg
MikeSteed
Posts: 11
Joined: Sat Aug 10, 2013 7:07 am

Re: Bug in FE Update 2 Date Command

Post by MikeSteed »

@Raziel

I too have a >NIL: after the command, but for some reason it always crashes on my system, rather than printing masses of garbage. That made it easier to track down the cause.

So far I've just been living with it. On the rare occasions when the crash occurs, I just tell the Reaper to kill the program. The only thing after the Date command in my Network-Startup is using SetClock to write the time to the hardware clock, which is pointless if Date fails. So no harm if Network-Startup never finishes.

As Colin noted, the bug is actually in timesync.library, so I suppose you could revert back to the previous version of the library. I never had a problem with that version.

Mike
User avatar
Raziel
Posts: 1170
Joined: Sat Jun 18, 2011 4:00 pm
Location: a dying planet

Re: Bug in FE Update 2 Date Command

Post by Raziel »

@MikeSteed

That would be a thing to test, thank you

I have lots of stuff starting after the network *ought* to be up.
Some personal AREXX scripts relying on working internet access.
I wrote myself a script that checks to *really* make sure internet is available before starting the remaining scripts.
It got way better after that check was in place, but i still have occasional boot freezes, where i can clearly see that the network is fine, but the "date" command never finishes and now that i know the reason and why it fails, it finally all adds up and makes sense...i blamed my scripts all the time :-)
People are dying.
Entire ecosystems are collapsing.
We are in the beginning of a mass extinction.
And all you can talk about is money and fairytales of eternal economic growth.
How dare you!
– Greta Thunberg
User avatar
nbache
Beta Tester
Beta Tester
Posts: 1714
Joined: Mon Dec 20, 2010 7:25 pm
Location: Copenhagen, Denmark
Contact:

Re: Bug in FE Update 2 Date Command

Post by nbache »

Raziel wrote: Fri Apr 16, 2021 10:09 amI wrote myself a script that checks to *really* make sure internet is available before starting the remaining scripts.
I'd be interested in knowing how you go about checking that. I have been trying various things in the past for the same reasons without finding the "ultimate" test.

Best regards,

Niels
User avatar
Raziel
Posts: 1170
Joined: Sat Jun 18, 2011 4:00 pm
Location: a dying planet

Re: Bug in FE Update 2 Date Command

Post by Raziel »

@nbache

Nothing ultimate here.
But it works for me so far.
It slows down WB boot a lot, compared to before, but at least I now *know* that the network is available and as I already wrote, it solved a bunch of boot problems i had...heck, even my cold boot with the case button works again. :-?

I typed below on the phone, so errors ahead.

Effectively I poll a ping until there are no more dropped tries. It doesnt matter if you reach out to your router order the www, no connection in either case unless network is running.

Code: Select all

/*
*/
v_network=0

DO WHILE v_network=0
        ADDRESS COMMAND 'ping -q -c 3 192.168.178.1 >T:ping.log'
        OPEN(pl,T:ping log','R')
        DO WHILE ~(EOF)
                v_ping=READLN(pl)
                IF INDEX(v_ping,'loss')>0 THEN DO
                        v_loss=COMPRESS(SUBWORD(v_ping,7,1),'%')
                        LEAVE
                END
        END
        IF v_loss=0 THEN
                v_network=1
        ADDRESS COMMAND 'wait 1'
END
EXIT 0
This script is started right after the AddNetinterface line, but is not "run" so everything after that has to wait for the script...and fir the network to be up.
People are dying.
Entire ecosystems are collapsing.
We are in the beginning of a mass extinction.
And all you can talk about is money and fairytales of eternal economic growth.
How dare you!
– Greta Thunberg
User avatar
nbache
Beta Tester
Beta Tester
Posts: 1714
Joined: Mon Dec 20, 2010 7:25 pm
Location: Copenhagen, Denmark
Contact:

Re: Bug in FE Update 2 Date Command

Post by nbache »

Yeah, okay, that would work, of course.

Although you won't (by pinging your router) know whether you have a connection further on into the internet, e.g. to an NTP server. I do my time syncing further down in Network-Startup, so I'd need that. Therefor I'd be pinging some remote server (wonder if my NTP server accepts pings?).

Another thing I'm not sure I would want is that it doesn't bail out after some fixed time or number of tries. Even more so for me if I were pinging someone else's server, I guess.

But otherwise, definitely a possibility, and if it works for your purposes ...

Best regards,

Niels
Post Reply