On Thu, Mar 2, 2017 at 4:59 PM, Harrison Tarr <harrison.t...@connexta.com> wrote: > I was able to get it to work by changing the ping command to “ping –n %d %s”. > I’m not sure why it skips the Network.isReachable; the process seems to be > running as an administrator anyway.
Now, should we make this the default on Windows? I'm a bit confused as it seems I was using cygwin. Anyone can help me? > > Thank you very much for your assistance and helping me to identify the > problem! > > -Harrison > > On 3/2/17, 1:44 PM, "Clebert Suconic" <clebert.suco...@gmail.com> wrote: > > Notice that the code will first try Network.isReachable(address) > before calling ping... > > the fact that you needed the ping, it means that Network.isReachable > wasn't called... > > > > > Notice: and this is a very important warning!!! > > I have had so restrictions about implementing the networkHealthCheck. > I even made a big warning on the docs... > > You must understand your Network topology. If you split brain your > network, there's no protection whatsoever... > > the fact that the network wasn't reachable without root access (which > is what ping.exe / ping gives you) it means that your check address > isn't on the same switch as you are. Be careful! > > > > > On Thu, Mar 2, 2017 at 3:23 PM, Harrison Tarr > <harrison.t...@connexta.com> wrote: > > This is the result of setting log lower: > > > > 12:17:51,179 | TRACE | e45-f6f221d82ada | > q.artemis.core.server.NetworkHealthCheck 343 | ogging.jboss-logging | > purePing on canonical address 192.30.252.153 > > 12:17:51,179 | DEBUG | e45-f6f221d82ada | > q.artemis.core.server.NetworkHealthCheck 365 | ogging.jboss-logging | > executing ping:: ping -c 1 -t 1 192.30.252.153 > > 12:17:51,210 | TRACE | e45-f6f221d82ada | > q.artemis.core.server.NetworkHealthCheck 381 | ogging.jboss-logging | Bad > parameter 192.30.252.153. > > 12:17:51,210 | WARN | e45-f6f221d82ada | > q.artemis.core.server.NetworkHealthCheck 206 | ogging.jboss-logging | Ping > Address /192.30.252.153 wasn't reacheable > > > > When I do “ping -c 1 -t 1 192.30.252.153” from the command line, I get > “Bad parameter 192.30.252.153.” just like in the logs. But when I do “ping > 192.30.252.153,” it works on the command line. Maybe when I experimented with > a different ping command in the broker.xml I didn’t do it right? What should > the command look like in the broker.xml for the command to be “ping > 192.30.252.153”? > > > > Harrison > > > > On 3/2/17, 1:09 PM, "Harrison Tarr" <harrison.t...@connexta.com> wrote: > > > > I am on 1.5.1. > > > > On 3/2/17, 1:05 PM, "Clebert Suconic" <clebert.suco...@gmail.com> > wrote: > > > > Are you using the latest (1.5.3) to debug this? > > > > Just to make sure I would be looking at the same git version I > need to look at. > > > > On Thu, Mar 2, 2017 at 3:01 PM, Harrison Tarr > > <harrison.t...@connexta.com> wrote: > > > I may very well be misunderstanding. > > > But I think my most pressing issue is that since the > NetworkHealthCheck is failing, my servers are not coming up at all, even when > they’re supposed to. I have tried to start my backup (without the ‘master’ > node running) and the backup does not start Artemis, because it can’t ping > any address. > > > > > > I’m going to investigate further, setting the log level to > trace, like Justin recommended. > > > > > > -Harrison > > > > > > On 3/2/17, 12:52 PM, "Clebert Suconic" > <clebert.suco...@gmail.com> wrote: > > > > > > Maybe you're misunderstanding how NetworkHealthCheck > works? > > > > > > > > > This is required only to make sure your server will not > run in case > > > you pull out the network cable. > > > > > > So, if you set the ping Address towards apache.org, your > server would > > > shutdown itself whenever your internet connection drops > or the > > > apache.org is shutdown. it will be pure ping. > > > > > > > > > > > > that's just to avoid a backup from activating because the > network > > > connection dropped. > > > > > > On Thu, Mar 2, 2017 at 2:41 PM, Harrison Tarr > > > <harrison.t...@connexta.com> wrote: > > > > Please find my broker.xml and backup broker.xml > attached. They work when I comment out the “<network-check-list>” but don’t > fail-over if I kill the master. > > > > > > > > Regards, > > > > > > > > Harrison Tarr > > > > > > > > On 3/2/17, 12:33 PM, "Clebert Suconic" > <clebert.suco...@gmail.com> wrote: > > > > > > > > It is possible to be a bug on windows... > > > > > > > > > > > > Can you share your settings? I will take a look. > > > > > > > > On Thu, Mar 2, 2017 at 12:29 PM, Harrison Tarr > > > > <harrison.t...@connexta.com> wrote: > > > > > Hello, > > > > > > > > > > I’m trying to configure HA using Artemis on a > Windows machine. My coworker was able to get the “network-check-list” to work > with a simple list of IPs. We are using the same broker.xml. When I use it, I > get an error in the log that says > > > > > “08:56:49,975 | WARN | ca9-d991d17b6f42 | > q.artemis.core.server.NetworkHealthCheck 206 | ogging.jboss-logging | Ping > Address /192.30.252.153 wasn't reacheable” > > > > > > > > > > The leading “/” in the IP seems suspicious to me. > I have tried using the default ping command, as well as editing it to be just > “ping.” I have also tried with a list or a single IP. I have also verified > that the addresses are reachable from the machine, just the ping fails. Could > there be some parsing error on Windows that is prepending a “/”? > > > > > > > > > > Regards, > > > > > > > > > > Harrison Tarr > > > > > > > > > > > > > > > > -- > > > > Clebert Suconic > > > > > > > > > > > > > > > > > > > > -- > > > Clebert Suconic > > > > > > > > > > > > > > -- > > Clebert Suconic > > > > > > > > > > > > -- > Clebert Suconic > > -- Clebert Suconic
