How to store the graph data on an external location?
Everytime when I update ROS I loose all graphs.
I even lost my graphs now in my cetral unit that only suffered from a netwarch reboot due upgrading (and thus down) next attached unit!
Any suggestions without the need to setup Dude or other external monitor programs.
Have you set up graphing to store statistic data to disk?
Off course, otherwise I couldn’t use it before…
But where is it stored on disk? In file it is not visible?
I have the same problem.. been running a main router on 3.30 for years. graphing resets every 12 months. Now with ros 5.6 on rb1100, the graphs reset after a reboot. yes it is set to store on disk.. ?
Hmm. I had router with 3.3 woith graphing turned on workign fine. Few days ago I upgraded to 5.4. I ahed to reboot several times in the process. Graphing data was kept (covers whole year).
However, i noticed that I cannot look grpahs in Winbox. It simply does not show that graphs exist. When I access graphing through web interface everything is fine.
After several reboots, graph for one interface showed up in winbox but others are still missing.
Any on this forum knows how to save graphs over reboots in 5.6+ versions? (I can recall my rb1000 lost these in 4.x versions too after a reboot…)
most likely the problem is the clock - when it reboots, its 1970 or so and it takes readings, and that probably resets the database. Maybe you could write some script to help test that theory. What time interval do you have graphing writing to disk?
Well, they are default. I set them in winbox and only through your remark I looked in terminal to see there is some time interval to be set. They are “5min” which I presume it default value.
Referring the clock, it certainly has strange behaviour. It takes some (administrators?) action to get a updated time tag in the log. As long as no changes are made to router it takes a long time before actions to be logged are tagged with the updated time. In the meantime all event are tagged by the default time which it makes it very hard at time to troubleshoot issues if you have a ’ feeling’ such issue might be triggered by an event from an other router down the pipeline. If this unit is rebooting all the time for unknown reason it is hard to use the log because the time tags are useless.
I can’t recall this was in early 3.x or 4.x versions the same. And in these days indeed the graphs did survive if I remember well. So you might be right, it might have something to do with the clock update after a reboot.