[FQuest Alert] SAMSON Server

[Bob]

The SAMSON Server has again locked up and is being tended to as I post this.

-Bob

[hobbes]

Time to send'm packing ...

[johnfl68]

Poor Samson. :QTcry:

[Bob]

The SAMSON server was returned to full service at 10:55 AM ET.

Our sincere apologies and you can believe that we are working very hard to determine the underlying issue so that SAMSON can be cured of the ailment that is obviously affecting him...

Thank you for your patience and cooperation during these trying days for SAMSON,
Bob

[Wassercrats]

I'd lay off Samson for six months to see if his replacement has the same problems. If it does, then you know it's not the machine. Servers are cheap, right? :safegrin:

[phppete]

Down again :(, I'm sending clients links to their dev sites and examples etc only to find the server is up and down more than a *****s knickers... come on guys this is a bit much and makes me look like a real lightweight when I can't even keep my site online.... perhaps you should dump it in the trash and buy a new server.

[Terra]

Pete, its not the hardware, but something else triggering a psuedo hang in the kernel... This is a completely different problem then the 0-Order hangs that we've dealt with...

The problem followed the drives (I wished it hadn't), when we did the FreakyFriday swap last night, and SERENITY's hardware has been very stable...

I can assure you the cause is being hunted down... I wished I could always wave my magic wand and fix everything in microseconds, but some bugs are tough and take time to figure out... This is one of those types of bugs, where over a hundred tasks are calling swap_out() and in turn is calling mmap() which in turn *tries* to grab the mmlist_lock spinlock which can't because another task has grabbed it and won't release the lock... Now the question is, what task is sitting on the lock (not known), what precursed it (known), and why it is not being released (not known) - ergo its a game of 'find the race'... In a highly active kernel, this is much much much more difficult than what it appears on the surface...

In conclusion, I don't get much sleep when there are problems like this...

--
Terra
sysAdmin
FutureQuest

[phppete]

I do appreciate the difficulty of the problem, I also realise we can't expect everything to run smoothly 100% of the time. It is just a little frustrating this end of obvious reasons, especially when I am showing clients work. Anyway I wish you luck in hunting down the problem :yeah: