[Lustre-discuss] 6 of 8 OST's unavailable after client failure

Ronald K Long rklong at usgs.gov
Fri Feb 11 07:47:54 PST 2011


We had a lustre client lose its nic connection to the lustre network.  The 
OSS's reported evicting this client.  However shortly after that 6 of the 
8 OST's reported issues with service threads and the lustre file system 
was unavailable on all of our clients.  Once we brought the client back on 
line the file system returned to normal.

I dont believe a client should be able to bring down the complete system. 
I have attached the messages log as well as the stack trace, if anyone 
could shed some light on this.

Thank you

Rocky 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.lustre.org/pipermail/lustre-discuss-lustre.org/attachments/20110211/074e1b12/attachment.htm>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: lustre_error.zip
Type: application/zip
Size: 1261939 bytes
Desc: not available
URL: <http://lists.lustre.org/pipermail/lustre-discuss-lustre.org/attachments/20110211/074e1b12/attachment.zip>


More information about the lustre-discuss mailing list