Hi all, I want to setup a RAID storage system, where i have two systems connected to it. the filesystems are mapped out to both connectors. I want the master host mount them read write, and the slave read only. in my fstab on the slave I have a line like the following: /dev/sdb1 /mount ext3 acl,noauto,user_xattr,nosuid,ro 0 0 so in man 5 fstab, it is written, that when the 6. field is 0, no filesystem check will be done at mount time. and in man mount, I read that, the nocheck parameter is the default, that means, that no filesystem checks should be performed when the partition is mounted. but when I mount the filesystem on the slave, I see the following messages in /var/log/messages: EXT3-fs: mounted filesystem with ordered data mode. EXT3-fs: INFO: recovery required on readonly filesystem. EXT3-fs: write access will be enabled during recovery. (fs/jbd/recovery.c, 255): journal_recover: JBD: recovery, exit status 0, recovered transactions 99610 to 100072 (fs/jbd/recovery.c, 257): journal_recover: JBD: Replayed 12498 and revoked 589/918 blocks kjournald starting. Commit interval 5 seconds EXT3-fs: recovery complete. since I test this, the master server had occassional problems with the filesystem, so he decided to mount these read-only, and I had to fsck it. I think the filesystem got destroyed because of the filesystem ckecks, while mounting it readonly on the second server. I googled around, and found a similar message from someone mounting a XFS file system. So I am not sure, whether this is a mount or a ext3 problem. my kernel is a 2.6.12.6-bigsmp, on a SuSE 10.1. Is there any other way to prevent the slave server from doing any filesystem checks? kind regards Sebastian
Running the following command on your slave server should do the trick: echo "AUTOFSCK_DEF_CHECK=\"no\"" >> /etc/sysconfig/autofsck ..Chuck..> -----Original Message----- > From: Sebastian Reitenbach [mailto:itlistuser at rapideye.de] > Sent: Tuesday, November 28, 2006 6:06 AM > To: ext3-users at redhat.com > Subject: how to prevent filesystem check > > Hi all, > > I want to setup a RAID storage system, where i have two > systems connected to it. the filesystems are mapped out to > both connectors. I want the master host mount them read > write, and the slave read only. > > in my fstab on the slave I have a line like the following: > /dev/sdb1 /mount ext3 acl,noauto,user_xattr,nosuid,ro 0 0 > > so in man 5 fstab, it is written, that when the 6. field is > 0, no filesystem check will be done at mount time. > > and in man mount, I read that, the nocheck parameter is the > default, that means, that no filesystem checks should be > performed when the partition is mounted. > > but when I mount the filesystem on the slave, I see the > following messages in /var/log/messages: > EXT3-fs: mounted filesystem with ordered data mode. > EXT3-fs: INFO: recovery required on readonly filesystem. > EXT3-fs: write access will be enabled during recovery. > (fs/jbd/recovery.c, 255): journal_recover: JBD: recovery, > exit status 0, recovered transactions 99610 to 100072 > (fs/jbd/recovery.c, 257): journal_recover: JBD: Replayed > 12498 and revoked > 589/918 blocks > kjournald starting. Commit interval 5 seconds > EXT3-fs: recovery complete. > > > since I test this, the master server had occassional problems > with the filesystem, so he decided to mount these read-only, > and I had to fsck it. > I think the filesystem got destroyed because of the > filesystem ckecks, while mounting it readonly on the second server. > > I googled around, and found a similar message from someone > mounting a XFS file system. So I am not sure, whether this is > a mount or a ext3 problem. > > my kernel is a 2.6.12.6-bigsmp, on a SuSE 10.1. > > Is there any other way to prevent the slave server from doing > any filesystem checks? > > > kind regards > Sebastian > > _______________________________________________ > Ext3-users mailing list > Ext3-users at redhat.com > https://www.redhat.com/mailman/listinfo/ext3-users >
On Nov 28, 2006 14:05 -0000, Sebastian Reitenbach wrote:> I want to setup a RAID storage system, where i have two systems connected to > it. the filesystems are mapped out to both connectors. I want the master host > mount them read write, and the slave read only.This is NOT possible with ext2 or ext3 and can result in filesystem corruption. Cheers, Andreas -- Andreas Dilger Principal Software Engineer Cluster File Systems, Inc.
use 'noload' option to mount the readonly ext3 filesystem on the slave host, so the journal will not be loaded. BTW, this kind of setting could have some cache-coherence problem. why not do it correct way by using some kind of network filesystem (NFS) or clustering filesystem (GFS,Lustre)? On 11/28/06, Sebastian Reitenbach <itlistuser at rapideye.de> wrote:> Hi all, > > I want to setup a RAID storage system, where i have two systems connected to > it. the filesystems are mapped out to both connectors. I want the master host > mount them read write, and the slave read only. > > in my fstab on the slave I have a line like the following: > /dev/sdb1 /mount ext3 acl,noauto,user_xattr,nosuid,ro 0 0 > > so in man 5 fstab, it is written, that when the 6. field is 0, no filesystem > check will be done at mount time. > > and in man mount, I read that, the nocheck parameter is the default, that > means, that no filesystem checks should be performed when the partition is > mounted. > > but when I mount the filesystem on the slave, I see the following messages > in /var/log/messages: > EXT3-fs: mounted filesystem with ordered data mode. > EXT3-fs: INFO: recovery required on readonly filesystem. > EXT3-fs: write access will be enabled during recovery. > (fs/jbd/recovery.c, 255): journal_recover: JBD: recovery, exit status 0, > recovered transactions 99610 to 100072 > (fs/jbd/recovery.c, 257): journal_recover: JBD: Replayed 12498 and revoked > 589/918 blocks > kjournald starting. Commit interval 5 seconds > EXT3-fs: recovery complete. > > > since I test this, the master server had occassional problems with the > filesystem, so he decided to mount these read-only, and I had to fsck it. > I think the filesystem got destroyed because of the filesystem ckecks, while > mounting it readonly on the second server. > > I googled around, and found a similar message from someone mounting a XFS file > system. So I am not sure, whether this is a mount or a ext3 problem. > > my kernel is a 2.6.12.6-bigsmp, on a SuSE 10.1. > > Is there any other way to prevent the slave server from doing any filesystem > checks? > > > kind regards > Sebastian > > _______________________________________________ > Ext3-users mailing list > Ext3-users at redhat.com > https://www.redhat.com/mailman/listinfo/ext3-users >
Hi, "Pengcheng Zou" <pengchengzou@gmail.com> wrote:> use 'noload' option to mount the readonly ext3 filesystem on the slave > host, so the journal will not be loaded.with the noload option, following output is shown at the mount command: mount: wrong fs type, bad option, bad superblock on /dev/sdb2, missing codepage or other error In some cases useful info is found in syslog - try dmesg | tail or so and in /var/log/messages: Nov 29 09:17:17 srv3 kernel: ext3: No journal on filesystem on sdb2 as this is a valid option, the same fs type, must be a bad superblock? is ther anything I can do about it?> > BTW, this kind of setting could have some cache-coherence problem. why > not do it correct way by using some kind of network filesystem (NFS) > or clustering filesystem (GFS,Lustre)? >yes, I need to take a look at these file systems. Sebastian