I am currently running 5.4 and are still experiencing the issue where the router becomes unresponsive after 3-5 days or more. A reboot resolves the issue.
Pinging from the router via the serial console gives the following error
[admin@akcr1] > ping 10.13.2.254
HOST SIZE TTL TIME STATUS
132 (No buffer space…
Any ideas?
Already lodged with mikrotik support along with supout.rif on the 03/06/2011 but no response, Ticket#2011060366000107
Stephen, I do not know what kind of help you expect to get on this topic.
We are working on this issue, but currently there are no results, as we are trying to reproduce the same problem without any luck.
Maybe you know about specific configuration required to get this problem repeated 100% of attempts.
Interms of replicating it, perhaps try setting up a pptp server on it, with a client trying to authenticate to it every few seconds and failing, do this continuously for 4-5 days. If that does not do it, try the same thing, but the connections failing via the gre portion of the connection failing to come up. ( We have a pptp proxy that a couple of clients connection via, that half dies on a regular basis)
Those are the only out of the ordinary things that go on, that I can think of.
Stephen,
we have PPTP/PPPoE/L2TP server in lab, clients are authenticated/refused, but there is no such error when default configuration is used.
Are there any other settings?
There should not be any unexpected error on the mikrotik side, except for when the mikrotik router becomes unresponsive and gives the no buffer space error.
The pptp proxy server I was refering to is not a mikrotik, Its www.mgix.com/pptpproxy, if that was causing any confusion. There a couple of connections via the proxy to the ROS pptp server.
I can tidy things up at my end, with regards to stopping the clients that are constantly trying to connect to the PPTP server and failing, and see if this prevents the issue, well it would technically just mean that the problem would take longer to appear.
I suggested this to MT earlier to make this a FAQ item or something prominent so that ROS users are aware of this problem:
If you use tarpit firewall actions in 5.x, it will fill up your route cache. The more connections run into tarpit, the sooner your cache fill fill up. The error message when talking about buffer space is really talking about route cache.
You can verify the status with
ip route cache print
There had been several threads and discussions about this with respect to versions 5.0-5.2 (most of them deleted now) … but it affects the whole 5.x line including all RC and Beta versions. Based on the information we received from support, this ROS behaviour is implemented in the kernel and as such is not under MT’s control. So this may not be posible to fix. You are better off replacing tarpits with drops
I may not be having the same issue, but the symptoms are the same. Running 5.7. Router does route, I can login, but I cannot do anything. It is basically non-responsive though I can open a terminal and navigate menus.
I am getting this error in RB493AH last 2 months. Modifications done before that happen first time :
upgraded from 4.xx to 5.xx
enabled PPTP server
Error happen once in 2 weeks maybe, there is only one pptp client, updated to 5.12. Also, can connect to to MAC address.
This is probably not related to problem but last thing in log is incoming connection tarpit, would be good if someone other can confirm if he is using tarpit.
I had this problem last few months and started to change some things in configuration to detect what is the problem. My router were freezing each 2-3 days for months and I GUESS I’ve found which is additional component to made this crash.
As first, I am still not sure is this is reason but that seems VPN server and tarpit together doing that, because, problem appear when I’ve enabled VPN server few months ago and now, since I’ve replaced all tarpit rules with drop router is stable for more than week. Will update this post if there is any change.
I should not talk After 9 days this crash again.
Well, in next test I will try not to use large address lists, one of my lists have more than 1000 items. Would help if someone who have same problem can describe router configuration …
RouterOS version 6.11 on RouterBoard 2011UAS.
Yesterday I had this problem twice. Both times reboot solved the problem. Interesting is that I was able to login using MAC address. So, layer 2 was working fine.
I don’t have Traffic Flow enabled.
I am using L2TP tunnels.
One thing that I noticed, and I don’t like it that I have 0.1% of bad blocks. Flash is 128 MB, so it is little more that 1 MB. Too much if you ask me. Does anybody know how much of bad blocks is not covered by warranty?
BTW, my configuration is very simple. Two WAN, one LAN, few queues, few L2TP servers, DHCP server, masquerade, few NAT rules…
Call me crazy, but 24 hours ago, on MikroTik’s website, there was no newer version than 6.11. I also checked for update with Winbox and there was no any.
After 9 days this crash again.