[erlang-questions] Curious: disparate net_ticktime's

[Tony Rogvall]

On 22 maj 2014, at 05:27, Scott Lystig Fritchie <fritchie@REDACTED> wrote:

> Ulf Wiger <ulf@REDACTED> wrote:
> 
> uw> This means that if there is no traffic between the nodes, there may
> uw> be problems if one node sends ticks at a lower frequency than that
> uw> expected by the other nodes.
> 
> AFAIK, this can also be tickled by a very large message sent between
> nodes and the network is "slow enough"(*).  The inet_drv driver may be
> busy transferring data in the middle of that message, but that busy-ness
> does not affect the ticktime ticker thingie.
> 

That should really be fixed. Why ticking if there is traffic going on between nodes?
Looking in dist_util.erl there is plenty of code trying to cope with that situation.
I guess if the send kernel buffers are large and the link is weak that may happend
anyway. Or ?

/Tony


> -Scott
> 
> (*) It might be that 1Gbit/sec Ethernet or 100Mbit/sec Ethernet is just
> too slow to transfer a 3 Gbyte Erlang term(**).  Or you've been bitten
> by a congested network that's dropping packets.  Or something else.
> 
> (**) This has never, ever, accidentally happened in a Riak cluster, oh,
> heavens, no.
> _______________________________________________
> erlang-questions mailing list
> erlang-questions@REDACTED
> http://erlang.org/mailman/listinfo/erlang-questions

"Installing applications can lead to corruption over time. Applications gradually write over each other's libraries, partial upgrades occur, user and system errors happen, and minute changes may be unnoticeable and difficult to fix"



-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://erlang.org/pipermail/erlang-questions/attachments/20140522/2cde8e5b/attachment.htm>