Hyper-V failover cluster - linux read only file system RRS feed

  • Question

  • I saw this post, which describes my problem fairly closely but did not find a solution.

    I have fail over clusters running Hyper-V on Windows 2008 R2 Server Core. Our storage is driven by a SAN connected via fiber channel to our hosts. The volumes for our VM's are VHD's inside of CSV's available to the cluster. The windows VM's perform well and have really no issues. Our linux VM's however seem to sporadically remount their file system as read-only (Far more than what is to be expected by file system errors). Sometimes I'll pull up the console on the linux VM's and it will just be sitting there "waiting for SDA devices" like it lost its storage. I haven't really been able to nail down if its only happening under heavy I/O load or not. Both Ubuntu 11.10+ and CentOS5/6 VM's will do this. I've noticed that this tends to happen only to VM's that we use SCSI drives instead of IDE. The IDE drives tend to be okay, although I really have no evidence to support that.

    Has anyone else seen this issue? Anyone else have any issues with running CentOS/Ubuntu on their hyper-v hosts?

    I'll try to post more details as I get them. I wanted to see if anyone else in the community is running CentOS/Ubuntu hosts and what their experiences have been.

    Thursday, July 19, 2012 6:51 PM