Disk Full.. runs out of space over time.. any idea?

[matthew]

I've have just experienced this problem with firmware 1.5.14 on WS-26-500-DC switch, and the same issue on the 12-250-DC, this is the reason a purchased a new 26-500-DC . New configuration.. after a while saving the configuration it throws this error. Is there a way of fixing without rebooting the switch?
this is the error..

Warning: file_put_contents(config.json.new): failed to open stream: No space left on device in /www/config.php on line 776
Warning: rename(config.json.new,config.json): No space left on device in /www/config.php on line 780

Press ENTER to confirm configuration change. If you do not hit enter within 60 seconds the switch will revert to the previous configuration.

mtdblock6 - jffs mount is not full on other switches only 812k used


jffs file list shows only 943k - where is the other data?


After another 2 months of using the switch, after the last reboot - the disk partition has started getting full again showing 40% used so far..

/dev/mtdblock6 4.3M 1.7M 2.6M 40% /jffs

after running du -h in /jffs directory .. I get 919k total used - these values don't match (files and partition usage of /jffs) and in a few months it will run out of space.. the only fix is to reboot the switch again disrupting services.

Any idea's on why this is doing this? I cleared out the log.. no difference.

[matthew]

I've done a reset ... factory as some articles say.. setup switch again. Over time it runs out of space. every time you save the config the partition fills up and eventually runs out of space... just depends on how many changes you make on your switch over time. I power on and off one of my backhauls during peak periods. I've seen this on multiple switches different models same firmware same bug.

[Flo]

We had the same issue about two weeks ago on a WS-26-400-AC switch running firmware 1.5.12.
I can confirm that every configuration change (Save/Apply) results in a slight loss of available jffs2 space that is not recovered.
The device affected had an uptime running of about 1.5 years with multiple configuration updates and recovered available jffs2 space after rebooting the device.
Maybe the default jffs2 filesystem has been formatted with less than-needed erase blocks or a running process blocks the outdated, obsolete old configuration: https://linux-mtd.infradead.narkive.com ... filesystem
Would be nice if this issue can be fixed with a firmware upgrade.