SmarttNet Technical Support
tech at smartt.com
Thu Nov 25 20:22:15 UTC 1999
Today I manually ran the news.daily.
Here are my observations:
For the first 3 minutes, innreport took about 95% is the cpu. When this news
server crashes, it's during the initial few processes that the expiration runs.
When this server dies, all the processes go to sleep , top shows 98% idle but
the load skyrockets to over 10.
Currently im on a 28 day remember. The history is 780 MB and the expire process
just took over 193 Mb. Even INND stays at 140 Mb.
I would love to track this annoying and recurring bug.
Any suggestions would be appreciated.
On Wed, 24 Nov 1999 21:40:35 GMT, tech at smartt.com (SmarttNet Technical Support)
>This is still continuing..
>Now I'm running cleanfeed. Is there a way I could filter these messages so that
>it does choke the news.daily process?
>On Fri, 08 Oct 1999 15:46:32 GMT, tech at smartt.com (SmarttNet Technical Support)
>>Here's a real mystery. As a note, this has been happening with INN 2.2 and
>>2.2.1 for awhile. This re-occurs ever 2 - 4 weeks.
>>When our servers crashes it is during the first 2 minutes of the expiration
>>process. And everytime it does, I find this in the next expiration process.
>> Duplicate message-id "[1DC40EC5FD017D9E366FEABECCF7D3FF]" in history
>> Duplicate message-id "[FD1E94A0AED9012C558B3DDBE3AC8352]" in history
>> Duplicate message-id "[B12EA6375D2D428C756E3D3C507F0803]" in history
>> Duplicate message-id "[4DC93D4F3C4665DB4030A190765128C2]" in history
>> Duplicate message-id "[1ED8EE0C6D537F68C08EA19C142746AF]" in history
>> Duplicate message-id "[0A16CC1E8BA246F8F6A82192AEA2DF10]" in history
>>It is more like 50 - 100+ lines of these errors. I have had the upstream (The
>>local TELco) try and trace them down without any luck.
>>Anyone seen anything like this?
More information about the inn-workers