Page 3 of 10

v1.0.6rc1

Posted: Wed Nov 19, 2014 3:05 pm
by sirhc
Sorry I forgot to post v1.0.5 but it was exactly the same as v1.0.5rc2, no changes

FIXES
- fixes port details for ports 13 & 14 in the WS-12-250A and WS-12-250B

Released 11/19/2014

v1.0.6rc3

Posted: Thu Nov 20, 2014 1:45 pm
by sirhc
1.0.6rc2 is a special version I am testing to solve a very remote issue which is why we skipped it.

FIXES
- fixes that when you do a factory default in the UI or CLI that it turns all the POE ports OFF.

EXPLANATION
Remember that normally we leave POE ports active during normal reboots from firmware upgrades or UI reboots so as to minimize tower down time. I just forgot to clarify to the programer that during a factory default you must turn them OFF. A total over sight on my part I am really sorry and frankly feel pretty stupid!

Released 11/20/2014

v1.0.7rc2

Posted: Mon Dec 01, 2014 7:00 pm
by sirhc
FIXES
- R/STP issue with multiple dumb switches, discovered by LRL
- R/STP issue when using Loop Protection and R/STP together

KNOWN ISSUES STILL WORKING ON
Rory thinks there may still be an issue that "sometimes" occurs when Loop Protect is used in conjunction with R/STP.
Rory did not have time to completely LAB this to his satisfaction but will complete the LAB tomorrow and we will post a newer version if needed.
However if you are using R/STP and Loop Protection is disabled then there is definitely no issue.

Released 12/1/2014

v1.0.7rc2 Continued

Posted: Wed Dec 03, 2014 12:37 pm
by Rory
I just wanted to add some clarification to the above information. The issue between STP and loop protection occurs infrequently, and only when there are multiple STP unaware devices bridging between STP aware devices (unless the STP timers are set to ridiculous values). Either way, when the issue occurs Loop Protection will disable the port. LP does this by setting the operational state of the port, not actually dropping the link. Because of that mechanism, STP is unaware that the port is actually down. STP then puts the port in forwarding mode, because the link is up and it is not receiving any BPDUs.

The end result of this is that LP continues to disable the port every three minutes until the switch is rebooted. This issue is at least mitigated in firmware v1.0.7rc2, but it still occurs very infrequently. I could only reproduce it once in many tests, unless I deliberately broke STP. If I configured the timers incorrectly I could reproduce the error every time.

So it *MAY* be advisable to disable Loop Protection if you must use spanning tree. At the very least, if you are unfamiliar with the various timers, hello, hold, etc. I would highly recommend brushing up on the documentation. The best route to take if you feel you need both technologies working immediately would be to ensure STP is properly configured. If you want to be safe just disable Loop Protection. STP does work properly without it.

Eric and I are working as hard as we can to iron this out. There have been some unexpected issues in implementing the fix, and I do not feel comfortable with greenlighting the newest release candidate until we square away these new roadblocks. Our opinion is that STP and Loop Protection should compliment each other and work well together to ensure the most robust network reliability.

Thanks for your patience!

v1.0.7rc6

Posted: Fri Dec 05, 2014 12:40 am
by sirhc
FIXES
- Finished fixing R/STP and Loop Protection as well as made them play nice with each other

- Fixed Ping Watch Dog

- Changed formatting of port bounce schedule dialog to be more intuitive

- Fixed several CLI issues including VLAN and LAG configuration

- Changed the way config is applied and saved in CLI


Released 12/4/2014

v1.0.7

Posted: Sat Dec 06, 2014 3:45 pm
by sirhc
FIXES
- traffic storms such as from a loop no longer kill the CPU
- loop protection notifies RSTP when enabling/disabling a port
- some R/STP configuration settings were not being applied until after reboot
- adjusted max fan speed to be at 85C which makes the switch quieter (our switch core chips are dye rated to 125C)
- fixed ping watchdog
- fixed problem with "show watchdogs" CLI command
- reduced cold boot time
- fixed allowed ranges for STP values
- ping watchdog should not attempt to email if SMTP is not enabled
- changed formatting of port bounce schedule dialog
- fixed several CLI problems
- changing config in CLI is now immediately applied when leaving config mode in the CLI
- added additional logging for STP and LACP events
- added logging for configuration changes

Released 12/6/2014

v1.0.8

Posted: Sun Dec 07, 2014 2:09 pm
by sirhc
FIXES
- we broke the MAC table in v1.0.7

EXPLANATION
None, just a dumb mistake, no excuse - SORRY :headb:

Released 12/7/2014

v1.0.9

Posted: Thu Dec 11, 2014 6:48 pm
by sirhc
FIXES
- fixed redirect if HTTPS is not default port

- fix ntp client bug

- fixed mac table filtering for entries on multiple ports (LAGS)

ENHANCEMENTS

- changes to format of log messages

- added SMTP alert for fan failure

- added loopback ability to switch utility (may add to UI at some point in the future)


Released 12/12/2014

v1.0.10rc5

Posted: Tue Dec 16, 2014 7:50 pm
by sirhc
FIXES
- tweaked TX/RX Limit syntax
- re-enable any disabled ports when Loop Protection is disabled
- log message tweaks
- DHCP client now always sets default route
- SMTP configuration no longer requires username/password
- can now manually configure the From address for email notifications
- disable HTTPS from CLI now works

ENHANCEMENTS
- added option for STP on LAGs
- can now jump between interfaces while configuring in CLI
- login password is now hashed on flash and config exports (prevents downgrading firmware below 1.0.10)

Released - 12/16/2014

v1.0.10rc6

Posted: Wed Dec 31, 2014 11:53 am
by sirhc
FIXES
- Fixed problem where a power failure would cause the switch to not re-power the POE ports on reboot

EXPLANATION
There was a flaw in our code that allows ports to normally stay powered up during a firmware upgrade or software reboot.

Even though this is an RC version PLEASE
upgrade to this version as previous versions will fail to properly power up ports after a power failure.
I am sorry for the inconvenience but you should upgrade switches in service with this version as soon as possible.

I could not push out a non RC version as we are getting ready to release firmware v1.1.0 with most of the features on the firmware roadmap very soon.

KNOWN ISSUES
If you factory default the switch you will need to console into it and set the password. My guess is this due to recently hashing the password.
NTP service still has issues, usually disable and then enable NTP in UI gets correct time.

I will ask Eric to push out a v1.0.10 final version with these fixes 1/2/2015.

I hope to release v1.1.0rc1 next week with most if not all the features on the firmware road map.

Released - 12/31/2014