graphs deleted after reboot on v.5.0 / slow NTP sync

I am using graphing set to store data on disk. After upgrade RouterOS to v.5.0, graphs are deleted 5 minutes after reboot (using RB750G and RB433). Maybe the problem is that clock is still not synchronized after 5 minutes (it takes 5 minutes and 20 seconds to synchronize using NTP Client and package ntp). So that 5 minutes after rebooting, system is generating graphs with clock set to year 2002 and probably deleting previously saved graphs from year 2011.

Is it a bug? I didn’t have any problems with graphs on v.4.xx.

no it isn’t a bug!
your job is ok
but you can change the timr of updating the graphs in “graphs”.
after any restarting.
you lost the previous graphs!
it is the rule of routers
good luck

I know only setting how often graphs are saved to disk (1 hour or 5 minutes). Where can I change time of updating graphs?

I’ve restarted routerOS v.4.xx many times and graphs were not lost (at most only last hour) - that’s because data were saved to disk every hour.

I just run into the same problem. The graphs disappeared.

print the values

[admin@MikroTik] > tool graphing print
   store-every: 5min
  page-refresh: 300

set store-every

[admin@MikroTik] > tool graphing set store-every=

StoreEvery ::= 24hours | 5min | hour

[admin@MikroTik] > tool graphing set store-every=5min

Please note I have set store-on-disk=yes for each resource/interface

[admin@MikroTik] > tool graphing interface print detail
Flags: X - disabled
 0   interface=all allow-address=0.0.0.0/0 store-on-disk=yes

From other side my NTP is fast and router has correct time right after the boot…
Regardless the old graphs are up for some 5 minutes (I guess StoreEvery time) and then they are overwritten instead of updated.

It would be very nice to have this graph continuous.

Regards,
Z

Problem still remains in v.5.1

plz see the picture
graphs02.JPG
graphs01.JPG

Yes, this allows to set the frequency of storing data to disk. I’m using 5 minutes on some routers and 1 hour on some other - but all of them delete data after reboot in RouterOS v. 5.0 and 5.1. That’s the problem (and was not a problem in v. 4.x).

I also suffer from disappearing graphs :wink: Should we file a bug report? I’m running RouterOS 5.1.


Edit:
Once the graphs lasted after reboot, but only once. Since i did more than one reboot i’m not satisfied with the results…

It really looks like NTP synchronization problems and data gets overwritten. Does anyone tested if changing store time to 1hr fixes the problem? NTP should be synced by that time…

seems like ROS should simply check whether current time is in the past compared to last graphs update. MT Staff?..

Hello,

I do not thing that this issue has anything to do with NTP sync.
My RB450G on 5.1 gets the correct time in few seconds - regardless the graphs last some 5 minutes after the reboot (I do have store-every=5min) and then they are re-setted.

Reagrds,
Z

Well, I did some tests with synchronizing NTP as fast as possible and the graphs stayed after reboot in 6 out of 12 tests. It looks like a random “feature”.

Hello,

unfortunately v5.2 does not solve this issue :frowning:

Regards,
Z

Anyone tried the new 5.4 and graphs? Do they still disappear?

I recently updated from v4.17 to v5.4 and still graphs resetted after each reboot of my 450G (NTP servers on internet, need pppoe interface to come up first).

Problem remains in 5.5 as well :frowning:

It seems that it is appearing only if the NTP package is installed - couold you confirm this idea?
I have uninstalled NTP package on some routers and they are remembering graphs after reboot now. But I don’t want to uninstall the package because I’d like to have an internal NTP time server.

AFAIR, SNTP client synchronizes much faster than NTP client…

Has anyone reported this bug/“feature” already?

Probably not.
Let’s make sure that this problem is appearing only if the NTP package is installed - reporting this fact would probably help to fix the problem.

If you do need NTP Server and you have to disable package NTP, you may use this workaround:

/ip firewall nat add action=dst-nat chain=dstnat comment="redirecting NTP" disabled=no dst-address-type=local dst-port=123 protocol=udp to-addresses=217.73.16.2

(where 217.73.16.2 is NTP server, you should use your favorite one)

If you would like to have exactly the same NTP server for the router and for clients as well, you may use scrpt like that:

:local ntp [:resolve sk.pool.ntp.org]
/system ntp client
set primary-ntp=$ntp
set secondary-ntp=$ntp
/ip firewall nat set to-address=$ntp [find comment="redirecting NTP"]

(where sk.pool.ntp.org is NTP server, you sholud use your favorite one)

I have scheduled this script to run every morning.