StarOS v2.01.0rc1-4528 ready for testing

[tony]

Do not apply this update unless you have easy access to the system in question. Always backup your configuration before upgrading.

Read the previous beta release notes before applying this update.

Release notes for v2.01.0rc1:
*) introduction of a new realtime traffic monitor called 'beacon'.
*) atheros wpa-client would not authenticate while bridged.
*) traffic monitor now shows kbytes/s as opposed to kbits/s

Release caveats:
*) none

[pppoe note]
There is now mppe support, along with mss clamping available to the pppoe server. enabling mppe clamping may have a negative performance impact on some systems. mss clamping must be enabled to use mppe.

As with all betas, please report your success using this release.

OEM versions of this release will be made available.

STARUTIL 1.11
Linux:
http://www.star-os.com/downloads/starutil-linux-1.11.tgz
Windows:
http://www.star-os.com/downloads/starutil-windows-1.11.zip

DESKTOP SYSTEMS
Desktop Server Edition: (32MB flash required)
http://www.star-os.com/downloads/oem-vnc/strs-2.01.0rc1-4528.iso
http://www.star-os.com/downloads/oem-vnc/strs-2.01.0rc1-4528.bin
http://www.star-os.com/downloads/oem-vnc/strs-2.01.0rc1-4528.raw

Desktop Router Edition: (16MB flash required)
http://www.star-os.com/downloads/oem-vnc/strr-2.01.0rc1-4528.iso
http://www.star-os.com/downloads/oem-vnc/strr-2.01.0rc1-4528.bin
http://www.star-os.com/downloads/oem-vnc/strr-2.01.0rc1-4528.raw

SPECIAL PURPOSE (SBC) SYSTEMS
Soekris & MODO Edition:
http://www.star-os.com/downloads/oem-vnc/strrs-2.01.0rc1-4528.iso
http://www.star-os.com/downloads/oem-vnc/strrs-2.01.0rc1-4528.bin
http://www.star-os.com/downloads/oem-vnc/strrs-2.01.0rc1-4528.raw

WRAP Edition:
http://www.star-os.com/downloads/oem-vnc/strrw-2.01.0rc1-4528.iso
http://www.star-os.com/downloads/oem-vnc/strrw-2.01.0rc1-4528.bin
http://www.star-os.com/downloads/oem-vnc/strrw-2.01.0rc1-4528.raw

Routerboard Edition:
http://www.star-os.com/downloads/oem-vnc/strrb-2.01.0rc1-4528.iso
http://www.star-os.com/downloads/oem-vnc/strrb-2.01.0rc1-4528.bin
http://www.star-os.com/downloads/oem-vnc/strrb-2.01.0rc1-4528.raw

CPE SYSTEMS
CPE Editions for v1.13.4b2 are available.

[djdodo]

Hi,why you change in this release all values e.g. in the DETAILED STATISTIC from kbites to KBYTES?

IMHO better is kbites...

[tony]

showing the values in kbits/s confused many users, and have had several requests to have it show kbytes/s instead.

Many of the stats shown using the older traffic monitor can be done via beacon, including the detailed statistics screen. (simply uncheck all display options except for proto)

[datawork]

:lol:

Really like the beacon option.

Six hours and no issues so far exept for the standard 733 reboot lockup.

Now have build installed on wrap, pci based atheros and Intersil based boards without issue.

Upgraded all boxes from 2.00.3.b4

Nice work guys

[flavius]

Hello Tony and Loonie,

After upgrade two boxes from b10 to rc1, 6 hours later
my ethernet card stop responding, like autosense problem.

Rebooting and all OK again. More 6 hours and ... bingo!

I downgrade to b10 and works fine.

How can I help? Some ideia?

Thanks,

Flavio
Brazil

[tony]

What hardware are you using? There have been no changes between the two releases in regards to ethernet support. If possible, can you verify that you are installing the correct edition for your platform, as that can have an impact on performance and ethernet reliability.

[datawork]

All the boxs we upgraded are running the router build if that helps.

[tony]

Thank you datawork.

[flavius]

Tony,

Thank you and datawork for yours answers!

I´m only using server version, b10 or rc1. b10 is ok.

Now my b10 is...
Server Uptime: 7 hours, 41 minutes and 44 seconds
...without problems.

My ethernet configuration (2 cards) and
more 1 DWL-AG530 (AR5213) using IRQ 10.

Not used card:
--------------------------------------------------------
8139too Fast Ethernet driver 0.9.26
PCI: Found IRQ 12 for device 00:0c.0
eth0: RealTek RTL8139 at 0xe080d000, 00:08:54:11:ff:6e, IRQ 12
eth0: Identified 8139 chip type 'RTL-8100B/8139D'

Problem card (with rc1):
--------------------------------------------------------
VIA Rhine Farmily Fast Ethernet Adapter Driver Ver. 4.11
Copyright (c) 2002 VIA Technologies, Inc.
PCI: Assigned IRQ 7 for device 00:12.0
eth1: VIA VT6102 Rhine II Fast Ethernet Adapter
eth1: MAC=00:0E:A6:BD:04:65 IO=0xb000 Mem=0xe081e000 IRQ=7

No IRQ conflicts. CF 20MB card. A7N8X-X. Sempron 2400+ 128 MB RAM

Exist some driver diference between router and server versions?

Any ideia? Thanks again.

Flavio
Brazil

[tony]

b10 and rc1 releases are identical outside of the changes posted in the release notes. There have been no foundation changes (drivers, etc.) between the releases.

For the desktop releases, the router version is the same as the server, except without web cache, or radius support. (smaller release for 16MB flashes). Hardware support is the same.

If you swap the cable & ips from your VIA network card, to the unused RTL8139, does the problem remain?

Can you post the filenames you downloaded for b10 and rc1?

Also, if you are able to provide a support file for both b10 and rc1, that will help.

Thanks!

[rasimoes]

Tony, i'm having problems with PPPoE on wireless interfaces... Look:

2004-11-08 19:33:04 Daemon.Info 200.208.105.33 pppoe-server[1164]: Sent PADT

2004-11-08 19:33:04 Daemon.Info 200.208.105.33 pppoe-server[1403]: Session 61 created for client 00:02:78:e6:b1:2a (1.1.1.61) on wpcm0 using Service-Name 'Life'

2004-11-08 19:33:04 Daemon.Info 200.208.105.33 pppd[1403]: Plugin /etc/ppp/plugins/rp-pppoe.so loaded.

2004-11-08 19:33:04 Daemon.Info 200.208.105.33 pppd[1403]: RP-PPPoE plugin version 3.3 compiled against pppd 2.4.2

2004-11-08 19:33:04 Daemon.Error 200.208.105.33 pppd[1403]: unrecognized option 'wpcm0'

2004-11-08 19:33:04 Daemon.Info 200.208.105.33 pppoe-server[1164]: Session 61 closed for client 00:02:78:e6:b1:2a
(1.1.1.61) on wpcm0

PS: Tried desktop and server editions...

[tony]

Thank you. This appears to be an isolated problem with using pppoe / pptp with pcmcia wireless cards, and will be resolved in the next release.

Thanks!

[rasimoes]

Thank you, glad to help! :wink:

[keith.yoder]

I recently installed the rc1 beta release on two wrap boards with two senao prism cards. Twice, my cpu usage has gone to 100% and stayed there for a few hours before coming back down. It could be conincidence - a user with a virus that came at the same time as my upgrade but I thought I'd post this to see if anyone else has had something simular. I'm going to downgrade one of the wraps to see if that helps. I was using the beta8a release before.

Keith Yoder

[lonnie]

Did you try the new bandwidth monitor? If it was someone with a virus and they were flooding you it would have shown them.

[patos]

Hi ,

I am runing on the latest version the VDS >. Virtual Tun .
Seems nice and efficient , but the Virual TUN have shaping capabilty , also .

HOw i can have shaped tunnels ? ....

[lonnie]

Use CBQ as you would with a real Ethernet device, using the vds1, vds2, etc names.

[gwinton]

Tony or Lonnie,

I been waithing to see if someone would explain VDS before I ask. But I have not seen this. SO here it goes. Can you explain VDS and give an example of what it would be used for.

Thanks and sorry for not knowing what this is.

[lonnie]

http://forums.star-os.com/viewtopic.php?t=3498&highlight=vds

This has become a discussion of VDS.

[kishvet]

Greetings,

Updated a WRAP board to this beta a few days ago and things went like they should - smooth. Like the new Beacon information. Decided to upgrade 4 servers and about a dozen wraps last night - not so smooth. The first server is our border router . Dual PIII with 2G of Memory. Star supplied DOM. 1 Intel Pro100 and 4 Realtek cards. No wireless. Uptime of 189 days with version 1.13.?. Upgraded but couldn't SSH into the box. Rebooted a few times and nothing. Figured the board booted into the bios screen and locked. Hooked up a monitor and keyboard and StarOS was up and running. Couldn't ping or pass traffic anywhere. X&%@##. After mentally running through my troubleshooting checklist, everything seemed to be configured properly and the same way it was before the upgrade. Finally figured out the new version somehow caused the renumbering of the Ethernet cards. What was previously known as ether1 was now ether5 and the rest were increased by one (Ether 2 became Ether1, etc). Of course now all of the connected devices are wrong as the interfaces took all of the configuration data with them to their new home. After rerouting all of the cat5 cables to their new homes, 3 of the interfaces are working like they should and two will not pass any data at all. The three that are working are plugged into switches and the 2 that are dead are plugged into other hosts (1 into a WRAP board and one into a Windows2003SBS). At least 15 reboots of the Star box and the 2 hosts as well as switching/reseating the cables/deleting the addresses and reinstalling/replacing the offending Ethernet cards and still no traffic. I figured it was some sort of an arp problem. I hooked a laptop directly to the two hosts and can pass traffic like they should. I gave up and was looking for a sledge hammer when all of a sudden the constant pings starting returning on one of the interfaces. Ten minutes later, the other one started to pass traffic and have been fine ever since.

Server number 2 with 6 Ethernet and 2 pcmcia radios also locked up on the upgrade. P4 Dell GSX260. Went on site and it was starting to load StarOS and stopped. Manual reboots and it stopped at the same place. On the reboots, the BIOS reports 3 instances of Plug and Play misconfigurations. Removed all Ethernet cards and reseated them and the BIOS errors disappeared. Had to burn an ISO of the new version and reinstall it on the DOM. I reconfigured and it worked fine.

Both of these units were multiple Ethernet interfaces. We have another Server that only has a single Ethernet and 4 pcmcia radios and it upgraded without a hitch. All WRAP boards also upgraded smoothly.

I have one more Server version that I need to upgrade but it has 1 radio and 4 Ethernet cards in it and I am scared to upgrade until I here from Lonnie and Tony. The machine is readily accessible and has a monitor and keyboard on it but I don't want to go through that hassle again.

Any ideas ?

Thanks.

Grant

[lonnie]

The new kernel adds some new devices and it seems that the devices are being scanned differently than the older 1.13 series.

I would say to recommend but be prepared to renumber the devices. I think the new features are worth the effort.

The issues you saw with the 10 minute delay was the connected untis requiring an ARP release. Once the arp was redone they came back up.

The radio will not change, but the 4 ethernets might very well change. Start with all RJ45 unplugged and start a ping to a machine on each segment and then start plugging cables in. By the process of elimination you will get them all connected. The first one will take the most number of tries.

There is not much we can do in this release. What we are working on is a name that does not change and it will be mapped to the MAC. Right now we use the Linux supplied name and that changes as per ordering.

[butchkemper]

If the ethernet cards are not the same chipset, then the relationship between the ethernet cards and the Linux device name will change if star-os changes the order in which it loads the Linux ethernet device drivers.

I had a similar problem with PCMCIA/PCI adapter cards when I installed a third card in the machine. The physical first and third PCI cards had a different chip set than the physical second card which caused physical card 2 to move from Linux device 2 to device 3. This was not expected and cause a brief time of head scratching.

Butch

[yurtesen]

Hi,

I dont know about previous versions but with this version the site survey doesnt work when the atheros card is already associated with an ap(or this is normal?). It only shows the AP it has been associated with. It would be nice to have the scan option when the card is in AP mode though. It would be fairly easy to put the card to station mode and scan the channels while scanning?

The PPTP connections do not work if I select show splash screen for pptp/pppoe users in hotspot. Only by unchecking it, things work fine. I think PPPoE connections are not effected. I am getting "You are outside of our service area" Now there was something about this in forum but without a solution posted... :(
http://forums.star-os.com/viewtopic.php?t=1943
Is there a fix for this? How can I get to see the splash screen? In the last beta version when I selected splash screen, I couldnt connect anywhere not even the noservice.htm came...
It would be nice if only pppoe users could see splash page and pptp users do not vice versa. If it is possible to seperate those 2?

OK the rest are not so important stuff but just thinking...If easy to implement then might be useful...

It would have been nice to have an option to be able to disable a wireless interface... (put it 'down' ?)

Possibility of disabling compression on PPTP/PPPoE etc. might be useful for people using boards which do not have the necessary cpu power, for example for people using WRAP boards. (I realized when I connect with pptp the compression is enabled unless I disable it from windows...)

Possibility to load defaults of some configuration files without resetting everything. For example wireless security, ip firewall etc.

Thanks,
Evren

[yurtesen]

the firewall script line

ldeny tcp from any to any 8080 in via $net

is twice in the default firewall script, maybe somebody should remove the extra one or there is a reason for it being twice?

[tony]

Thank you for the heads up. The duplicate entry has been removed.

[yurtesen]

I also realized in proxy settings that the default proxy size is 100mbyte even though I had 47mbyte for proxy in hard drive.
It is a little bit weird because the size should be set according to available memory and hard drive space. Probably it should be calculated at every boot and updated automatically. (to detect memory size changes etc.)


http://www.squid-cache.org/Doc/FAQ/FAQ-8.html#ss8.11
8.11 How much memory do I need in my Squid server?
As a rule of thumb on Squid uses approximately 10 MB of RAM per GB of the total of all cache_dirs (more on 64 bit servers such as Alpha), plus your cache_mem setting and about an additional 10-20MB. It is recommended to have at least twice this amount of physical RAM available on your Squid server. For a more detailed discussion on Squid's memory usage see the sections above.

The recommended extra RAM besides what is used by Squid is used by the operating system to improve disk I/O performance and by other applications or services running on the server. This will be true even of a server which runs Squid as the only tcp service, since there is a minimum level of memory needed for process management, logging, and other OS level routines.

If you have a low memory server, and a large disk, then you will not necessarily be able to use all the disk space, since as the cache fills the memory available will be insufficient, forcing Squid to swap out memory and affecting performance. A very large cache_dir total and insufficient physical RAM + Swap could cause Squid to stop functioning completely. The solution for larger caches is to get more physical RAM; allocating more to Squid via cache_mem will not help.

[tony]

The proxy is normally used on hard drives larger that 1GB in size. The options you mention are designed to be user tuned to their needs. The defaults should work well for most.

You mention 47MB for proxy. What kind of device are you using for storage? I would not recommend any solid-state device for the proxy as they have a limited number of times they can be written too before going bad. Normally this is in the thousands of times, however with proxy servers, this can add up quickly.

[yurtesen]

No, I am just stress testing staros in microsoft virtualpc :p so I made a 64mbyte hard drive for proxy+radius and gave about 16mbyte to radius...

I was just thinking about that 100mbyte thing because as far as I can remember, it is the default of squid. I mean if it is not set dynamically then somebody who put 10gbyte hard drive will only use 100mbyte of it...
and it says that we shouldnt mess with those in the conf file example :p

[tony]

The proxy cache_dir can be adjusted to allow for a larger overall cache size without worry. You may need to re-format your proxy drive if you change the levels.

Thanks!

[georgew]

The new kernel adds some new devices and it seems that the devices are being scanned differently than the older 1.13 series.

I would say to recommend but be prepared to renumber the devices. I think the new features are worth the effort.

My difficult to get to systems are WRAP boards.

Will a dual ethernet wrap with one or two atheros cards in it have any interface renumbering problems?

I suppose I can get in wirelessly if I have to renumber the ethernet.

Otherwise it's a 200' climb. I only plugged ethernet into the first port, but if that changes...

[lonnie]

The WRAP board will not rename any devices. It will be safe to do the update and you can expect to have the same settings as before.

[georgew]

What about the other wireless settings like speed and mode?

If I have it locked at a transmit rate of 11 mbps and the operating mode set to auto, with both cards in ap mode. One of these links is the backhaul. those settings should be ok,
but I probably should be ready to drive out there anyway.

So many features, so little time....

[lonnie]

The WRAP board will update fine.

[kishvet]

The WRAP board will update fine.

We have not had ANY problems with updating WRAP boards. As long as I know what to expect, we can live with the updating of the servers. Thanks.