What’s new in 5.4 (2011-May-27 13:18):
*) webfig - do not try to open many windows
if first open was blocked by browser;
*) RB4xx ether1 port flapping fixed;
Happy Birthday, Normis! 
hello master.
how to make group manually? for example, i want grouping for users that allowed for manipulating on queue menu, other menus are disallowed to access except queue menu.
please help me. 
unfortunately, you cannot
Hello,
Thank you for the new release.
I am sorry to be the first to complain but webfig still behaves very strange.
After a correct login it appears:
ERROR: Router has been disconnected
Reconnect
and it does it many times… but if I simply reload the page it shows the correct webfig interface.
RB450G ROSv5.4 Google Chrome
Also I have noticed that the graphs still disappear after some 5 minutes after the reboot causing the lost of all previous traffic history.
Thank you.
Regards,
Z
may you tell us more about this problem ???
After upgrade to 5.2 I´m having some issues with auto-negotiation.
*) RB4xx ether1 port flapping fixed;
thanks master 
,
i hope for next release mikrotik team will add my feature request “create group manually” into MT ROS.
Chupaka - you can do that now, in webfig look at the skin designer
yogii, open webfi and click “design skin”. remove all unneeded elements, and save with a new name, then open “system users” and make a new group that has access only to the new skin.
all should be fixed
thats right.. , thank you mr. normis. keep spirit for our ROS.
for mr. normis.
can i cofigure skin via winbox or terminal? i’ve search there is no skin via terminal.
i understand, skin feature just works on webfig right. thanks a lot MT team
yes, webfig can be customized and also translated now. this feature will only be available for webfig, but it has most of the features of winbox, it’s almost identical but works faster
then you, probably, need new ‘webfig’ security policy - to allow access via webfig and disallow access via webbox… because webbox is still available: http://router_ip/cfg . security hole?
hello master,
in mikrotik wiki ping tools page, site says packet size value is 64 bytes by default.
but its different when i test directly, ping size is 56 bytes?
if no response is received within 1000ms, ping will show as "timed out"
would mikrotik team change default value 1000ms to 2000ms or more? it will be difficult for me to monitor between my connection and my isp, i cannt distinguish actually my connection is time out or ping response more than 1000ms.
webfig skin is not a security measure, it’s a convenience feature to get rid of unneeded stuff
Hi guys,
TCP window scaling for outbound traffic is still not working well. It still not increases as it does without MT router.
Due to this throughput on high latency networks are very bad. It was partially solved in 5.3 pre-release, but it looks like its a bit worse in 5.4.
Without MT router: window size is 15083 (throughput is ~50Mbps)
02:57:28.472021 IP debian.41439 > *******.www: Flags [.], ack 15162677, win 15083, options [nop,nop,TS val 229785438 ecr 2217274565,nop,nop,sack 3 {16777197:16777597}{16772853:16775749}{16764165:16769957}], length 0
MT RB750G router (bridged interface nothing special): window size is 753 (throughput is ~1.5Mbps)
03:00:19.428143 IP debian.36092 > *******.www: Flags [.], ack 16660689, win 753, options [nop,nop,TS val 229828178 ecr 299801442], length 0
*) fixed RB450G, RB750G switch chip slow Ethernet problem;
Problem with Queues.
We upgraded a core router from 4.x to 5.4.. The router (HP ML310) has some simple queues to smooth trafic flows.
Soon after the upgrade we had a number of reports of specific web sites being offline or very slow loading, for example http://www.monarch.co.uk
After some investigation we found that Http traffic from this site would not pass reliably through Any sort of simple queue, even if that queue ha unlimmited bandwidth set. No other traffic seemed to be affected. Turning all queues off solved the problem and allow normal http access to that site.
System was downgraded to 4.17 and queues reinstated, problem is resolved.
Anyone see this or has more infomation. Whats new with queues in the linux?
Same here with queue trees. I have an “http/https” child queue in queue trees, and for the first time I’m seeing dropped packets in the queue. All other 7 queues have 0 dropped.
Problem with pings when advanced data rates enabled instead legacy. Support contacted with this:
I tried to: Enable advanced data rates instead legacy but pings to the
clients become very unstable, 2.4GHz or 5.6GHz, doesn’t matter.I saw this problem: When there is some clients with last activity values above 10…
ping reaches 30-50ms but when I disconnect that client manually, pings
are stable again. This problem occurs on every router, RBs and x86. Please, solve this.I expected to see: Everything become more effective than legacy rates but not.