View Full Version : StarV3 v1.2.1b build 2321 is ready for testing
The BETA has been released, and is available on http://files.star-os.com/
While the change log is not complete at the moment, there are some changes listed on this thread http://forums.star-os.com/showthread.php?t=6714.
Please do not upgrade your system unless you have access to it via the ethernet, in case you need to downgrade.
There are some known issues with this release, which include:
*) Transmit power may overdrive the card, if set too high. Symptoms include decreased signal quality at the higher rates. Set to 12 for good results. (note: only effects cloaking 2x and 4x modes)
*) RIP and OSPF may be restarted when a dhcp client rebinds it's IP, even if it has not been changed. (can occur if the client disconnects from the AP)
Some operational differences include (mainly related to regulatory requirements):
*) Atheros channel and country code list no longer contains an 'All Channel' setting (## and #!).
*) U1 country code (US + FCC3) no longer contains the FCC3 channels, however country UZ can be used in it's place for systems that require them.
For best operation, please ensure your client and AP use the same country code.
The website, and change logs will be completed over the next few days.
keith.yoder
06-24-2007, 02:05 PM
I guess I'll be the first one to post success here. I upgraded a WAR2 board, one that I have ethernet access. I haven't had time to try all the fancy new features yet. When I first rebooted the board one link that had signal of about -68 had dropped to -88. I quickly changed the Tx power setting 12, per Tony's post. That brought my signal back to -77. I get as good of troughput as before. This is a 42km link with 4x cloaking. I can get around 900kbps according to the Star-OS troughput test.
I assume that if you tried this upgrade on a more marginal link, it might not associate after rebooting unless you changed the power setting before rebooting.
Thank you for the update.
Does this upgrade work on a w1 board? I tried flashing one and it said no compatible firmware found in the provided package after uploading the .pkg file. It currently has build 2083 on it and is one the the units I recently received. I used Utilstar to upload, it went with no issues.
bobbyc
06-24-2007, 04:52 PM
With the previous beta build, I noticed a problem with newer atheros cards (such as the compex cards) where the rx sensitivity was poor. Older cards such as CM9 (5004?) didn't have this problem.
So far this new beta seems to have solved that problem.
Bob C
keith.yoder
06-24-2007, 05:20 PM
I think this was talked about in the previous beta thread but that was a long time ago. WEP doesn't seem to work from the 1.1.X branch to this 1.2.X branch. I had one link that wouldn't pass traffic until I deativated WEP. Maybe that should be in the release caveats :)
Quite a bit of the testing has gone into the WPA support and compatibility. I will revisit the WEP support, and ensure it is in good working order in the upcoming BETA.
Yes, the WAR-1 release is available, just make sure you install the MIPS-COMPEX release, and you should be good to go.
Keep the reports coming in.
keith.yoder
06-24-2007, 06:01 PM
Okay, maybe it wasn't the wep.
I have a short (2 or 3km) link. I get -80 signal, both in 1.1.13 and 1.2.1b. It will pass traffic sometimes but often it goes in to a state where the Quality indicator shows 65 (it is always 65) and it keeps scanning through the channels. It shows the Signal strength, etc, but doesn't appear to be connected.
In this case the client radio is the one I upgraded. Any ideas?
Ensure your transmit power is not set too high (or set to def). Set to to 10-12 and see if it helps.
rbolduc
06-24-2007, 07:23 PM
Well I upgraded to this beta on my personal CPE (WRAP) and lost connection to the AP Have set the power to 12 and when it did connect the AP showed a Signal of -89 where it used to be -75, I increased in steps all the way to 18 where I could actually use the connection with a signal of -83 and the quality was all over the place, I have downgraded to 1.1.11-1741 and set the power back to 22 where I am now able to use the connection back at full speeds and the quality is floating around 93-100..
The CPE is a Wrap with 2 WLM54G cards 18volt PS, 30' Cat5 Running G-Mode Only, lock rate at 12mb, x2 cloaking, Short Preamble, Super G Features
AP War 4 1.1.13-2080 WLM54G, Same settings as above 24volt PS..
Reed
i20access
06-24-2007, 08:11 PM
I have noticed that even though I have my client's data rates set to "6" in g mode the AP displays them as "9". Not sure if this is a bug or if I have a wrong setting.
Thank you
Which systems are using 1.2.1b, the AP or Client?
I have noticed that even though I have my client's data rates set to "6" in g mode the AP displays them as "9". Not sure if this is a bug or if I have a wrong setting.
Thank you
keith.yoder
06-24-2007, 08:19 PM
I switched my AP and Client radios (it is a PTP link). Now 1.2.1b is my AP and 1.1.13 is my cliente and everything is working (though I've lost about 4 or 5 db of signal).
I've been trying to play with the SNMP settings. I can get all standard information (InOctets, OutOctets, etc) but I can't figure out what OIDs to use for the 802.11 stuff. I'm probably showing my ignorance here, I'm no SNMP expert. Anyone have an example of a snmpget command they'd like to share?
i20access
06-24-2007, 08:24 PM
Which systems are using 1.2.1b, the AP or Client?
Both, and even some client systems are still 1.1.13 and showing as a data rate of 9 at the AP. Is there something I am doing wrong?
Also, at 2x cloaking do I use 12 as the TX rate on the AP & Client? It is seeming to give me MUCH more horrible signals.
Thanks,
Download and install our mib files from the Release-BETA download folder, and perform the following:
snmpwalk -v 2c -c public -m VLMT-EXP-802DOT11EXT-MIB 192.168.1.1 vlmt802dot11ext
The transmit power issues in cloaking modes will be resolved in the upcoming release. If you are locking your rate to 6 or 9M, you can safely set your txpower to 17.
i20access, to clarify the rate issue you are seeing, can you confirm that:
1. Your clients are set to 6M for a rate
2. Your AP is showing 9M (under the rx rate) in the association screen, or is it showing 9M under the tx rate? The rx rate represent the rate the client is sending to the AP, and the tx rate represents the rate the AP is sending to the client.
i20access
06-24-2007, 08:38 PM
The transmit power issues in cloaking modes will be resolved in the upcoming release. If you are locking your rate to 6 or 9M, you can safely set your txpower to 17.
Well, when I set my TX rate at the towers in question to 17 - I am unable to SSH to them from the rest of my network, only locally through the shell on the AP. When I set them to 12, my quality basically sucks but I am able to ssh to them as I normally would.
i20access, to clarify the rate issue you are seeing, can you confirm that:
1. Your clients are set to 6M for a rate
Yes, they are all set to 6 under TX rate.
2. Your AP is showing 9M (under the rx rate) in the association screen, or is it showing 9M under the tx rate? The rx rate represent the rate the client is sending to the AP, and the tx rate represents the rate the AP is sending to the client.
I understand which column is which, my AP TX rate it set (currently) at 18. It displays 18 in that column (the TX column). The AP is displaying "9" in the RX column. Would you like me to IPMAP a public IP to the AP in question so you may login and see this problem first hand??
Thanks Tony,
No, that shouldn't be necessary. Your system is currently using 802.11g, with 2x cloaking. This should be enough for us to duplicate the issue.
If it does show the wrong rate, it would be 802.11g specific (802.11a rates are displayed correctly), and that it would be cosmetic only.
Thanks!
i20access
06-24-2007, 08:45 PM
No, that shouldn't be necessary. Your system is currently using 802.11g, with 2x cloaking. This should be enough for us to duplicate the issue.
If it does show the wrong rate, it would be 802.11g specific (802.11a rates are displayed correctly), and that it would be cosmetic only.
Thanks!
Please also address my other issue with txpower of 17 (with a rate of 6) not allowing me to access the towers remotely, and only through the ssh from the AP.
lonnie
06-25-2007, 12:05 AM
It sounds like routing has not kicked in, Mark. Can you stop and restart RIP?
mimbach
06-25-2007, 02:07 AM
Guys,
I am having an odd problem.
Three towers the middle tower has a single war2 with 2 cm9's.
If I make any changes and drop the connection from tower three to tower 2 (the middle tower). Routing through the middle tower is dead until I reboot the radio on the middle tower. No other changes are made. Just the reboot. Connection tracking is disabled on all wars.
I also disabled olsr and setup static routes. Same result.
Thanks,
Mimbach
lonnie
06-25-2007, 08:28 AM
It definitely sounds like our activate script has messed up routing.
When you activate changes, but before you reboot - can you go to advanced->system reports->view system information and see what routes are listed?
I noticed you disabled OLSR, etc. but are you using RIP, OSPF or DHCP client on any of your towers?
Also, can you ping Tower 3, from Tower 2 after an activate?
i20access
06-25-2007, 08:54 AM
It sounds like routing has not kicked in, Mark. Can you stop and restart RIP?
I don't use RIP, OLSR is running on the AP and the towers (as clients) coming off of it are statically routed.
Thanks
mimbach
06-25-2007, 09:43 AM
In my case I don't use dhcp and all routing protocols are disabled.
There is a new release available.