Hi again! The problem seems to be filesystem related. I changed the filesystem on the external array to reiserfs and finally things work realy fine. My customer currently does some heavy load testing und right now everything works realy well.
Thanks for all the help!!! -----Urspr�ngliche Nachricht----- Von: Mohamed Badri [mailto:[EMAIL PROTECTED] Gesendet: Donnerstag, 10. Februar 2005 17:10 An: [EMAIL PROTECTED] Betreff: [Spamverdacht] Re: [gentoo-server] AW: [Spamverdacht] Re: [gentoo-server] Problems with HP MSA 500 Storage If vg is disabled he can't, because lvm2 devices are not created until vg is activated. any lvm2 specific errors ? Did you ask hp, if they support this controller with 2.6 kernel ? perhabs it's an issue with controller module in 2.6 kernel. Sorry but I can't help, I never see this type of architecture with ext2. In this case it is common to use gfs if the 2 hosts are active or any over journalized filesystems ( ext3, reiser ... ) But I think there is no link with filesystem as errors is sent by controller kernel module. Selon "Witschel, Matthias, Infraserv-Hoechst/DE" <[EMAIL PROTECTED]>: > Hi! > > I have heartbeat set up to control the entire volumegroup containing th > msa500 volume. It's disabled at boot time on both hosts via boot.local and > activated and mounted by heartbeat resource scripts. The only step to further > disable any unwanted contact to the filesystem is to take the ext2 kernel > support out of the kernel to a module, that would additionally by loaded by > heartbeat. Any idea if its possible for the standby host to write to the > volume if the vg is disabled? > > -----Urspr�ngliche Nachricht----- > Von: Mohamed Badri [mailto:[EMAIL PROTECTED] > Gesendet: Donnerstag, 10. Februar 2005 16:08 > An: [EMAIL PROTECTED] > Betreff: [Spamverdacht] Re: [gentoo-server] Problems with HP MSA 500 > Storage > > > Hi, > > did the msa500 lvm volume is mounted on each hosts at the same time ? > or do you have an heartbeat software that mount the volume after the master > host failed ? > > This problem can happen if two hosts try to write in the ext2 filesystem at > the > same time, as ext2 cant handle this case.
