3d downtime reported - why?
Hello, I have noticed that my probe data on the atlas.ripe.net-website reports at downtime of over 3 days: 2011-10-25 04:17:20 UTC to 2011-10-28 08:04:53 UTC. In fact, my connection was not down - maybe a few minutes (that seems to happen regularly), but everythin longer we would have noticed. We use this internet connection most of the day. My probe ID is not unter 500 but over 1000. Is this an error in the database, a problem with the reporting system or any other known bug? Did anyone else experience something like this? Or is there some issue I can resolve locally? Thank you Daniel AJ -- My twitter feed: @newstik http://twitter.com/newstik
On Wed, 02 Nov 2011 02:42:53 -0300 "Daniel AJ Sokolov (lists)" <listclient@sokolov.eu.org> wrote:
In fact, my connection was not down - maybe a few minutes (that seems to happen regularly), but everythin longer we would have noticed. We use this internet connection most of the day.
I saw similar issue previously. We had network problems some weeks ago and probe didn't get it's status up again before I gave it power-reset. I'm quite sure there is a problem in probe software. -- Tuomo Soini <tis@foobar.fi> Foobar Linux services +358 40 5240030 Foobar Oy <http://foobar.fi/>
On Wed, November 2, 2011 6:46 am, Tuomo Soini wrote:
On Wed, 02 Nov 2011 02:42:53 -0300 "Daniel AJ Sokolov (lists)" <listclient@sokolov.eu.org> wrote:
In fact, my connection was not down - maybe a few minutes (that seems to happen regularly), but everythin longer we would have noticed. We use this internet connection most of the day.
I'm quite sure there is a problem in probe software.
On that note, it would be really useful if the reason for the disconnect was shown. Probe restart/software upgrade disconnects are not as worrying (for me at least) as connectivity-related disconnects. If the probe sent a connection count since boot when making a connection, then I would be able to see which disconnects I can ignore and which need investigation. Michael
On 11/2/11 9:15 , Michael Procter wrote:
If the probe sent a connection count since boot when making a connection, then I would be able to see which disconnects I can ignore and which need investigation. Michael
With the current firmware, probes report their uptime on a regular basis. However that information is currently not processed by the back-end system. Probes also continue doing measurements if there is no network connectivity. However that doesn't show up in the graphs. It should however show up in the logs. If you see missing packets in the log files then there was probably some kind of network problem otherwise, it is probably a probe problem. Philip Homburg
On 11/2/11 6:42 , Daniel AJ Sokolov (lists) wrote:
I have noticed that my probe data on the atlas.ripe.net-website reports at downtime of over 3 days: 2011-10-25 04:17:20 UTC to 2011-10-28 08:04:53 UTC.
In fact, my connection was not down - maybe a few minutes (that seems to happen regularly), but everythin longer we would have noticed. We use this internet connection most of the day.
My probe ID is not unter 500 but over 1000.
Is this an error in the database, a problem with the reporting system or any other known bug?
We found that within the back-end, an update message got lost. When the connection to the controller came up, the fact that this happened was not recorded in the database. So the system reports the probe as being down during that connection. As far as I can tell, no data was lost. Philip Homburg
participants (4)
-
Daniel AJ Sokolov (lists) -
Michael Procter -
Philip Homburg -
Tuomo Soini