Re: Mirrormaker consumption slowness

Thanks Steve for your tips. Yes, we found many sacks in packet sequence of problematic connections and observed there was intermittent network jitter in between. That explained the behavior seen in our setup. Regards, Jeff On 12/7/17, 7:45 AM, "Steve

Re: Mirrormaker consumption slowness

This kind of sounds to me like there’s packet loss somewhere and TCP is closing the window to try to limit congestion. But from the snippets you posted, I didn’t see any sacks in the tcpdump output. If there *are* sacks, that’d be a strong indicator of loss somewhere, whether it’s in the

Re: Mirrormaker consumption slowness

Mirror mare is placed to close to target and send/receive buffer size set to 10MB which is the result of bandwidth-delay product. OS level tcp buffer has also been increased to 16MB max On Wed, 6 Dec 2017 at 15:19 Jan Filipiak wrote: > Hi, > > two questions. Is your

Re: Mirrormaker consumption slowness

Hi, two questions. Is your MirrorMaker collocated with the source or the target? what are the send and receive buffer sizes on the connections that do span across WAN? Hope we can get you some help. Best jan On 06.12.2017 14:36, Xu, Zhaohui wrote: Any update on this issue? We also run

Re: Mirrormaker consumption slowness

Any update on this issue? We also run into similar situation recently. The mirrormaker is leveraged to replicate messages between clusters in different dc. But sometimes a portion of partitions are with high consumer lag and tcpdump also shows similar packet delivery pattern. The behavior is

Re: Mirrormaker consumption slowness

Hi, any pointer will be highly appreciated On Thu, 30 Nov 2017 at 14:56 tao xiao wrote: > Hi There, > > > > We are running into a weird situation when using Mirrormaker to replicate > messages between Kafka clusters across datacenter and reach you for help in > case you

Mirrormaker consumption slowness

Hi There, We are running into a weird situation when using Mirrormaker to replicate messages between Kafka clusters across datacenter and reach you for help in case you also encountered this kind of problem before or have some insights in this kind of issue. Here is the scenario. We have