e2fsprogs: update mkfs defaults

The forced fsck often comes at unexpected and inopportune moments, and even enterprise customers are often caught by surprise when this happens. Because a filesystem with an error condition will be marked as requiring fsck anyway, I submit that the time-based and mount-based checks are not particularly useful, and that administrators can schedule fscks on their own time, or tune2fs the enforced intervals if they so choose. Patch #1 disables the intervals by default, and I've added a new mkfs option (-C) to turn on the old behavior of random, unexpected, time-consuming fscks at boot time. ;)

Signed-off-by Eric Sandeen <[email protected]>

---

lib/ext2fs/ext2fs.h +1 -0

index a204eb7..37584aa 100644
@@ -184,6 +184,7 @@ typedef struct ext2_file *ext2_file_t;
 #define EXT2_FLAG_64BITS		0x20000
 #define EXT2_FLAG_PRINT_PROGRESS	0x40000
 #define EXT2_FLAG_DIRECT_IO		0x80000
+#define EXT2_FLAG_MOUNT_FSCK_CHECK	0x100000
 
 /*
  * Special flag in the ext2 inode i_flag field that means that this is

lib/ext2fs/initialize.c +4 -2

index 7c38975..ff63a92 100644
@@ -155,7 +155,6 @@ errcode_t ext2fs_initialize(const char *name, int flags,
 	set_field(s_log_block_size, 0);	/* default blocksize: 1024 bytes */
 	set_field(s_log_frag_size, 0); /* default fragsize: 1024 bytes */
 	set_field(s_first_data_block, super->s_log_block_size ? 0 : 1);
-	set_field(s_max_mnt_count, EXT2_DFL_MAX_MNT_COUNT);
 	set_field(s_errors, EXT2_ERRORS_DEFAULT);
 	set_field(s_feature_compat, 0);
 	set_field(s_feature_incompat, 0);

@@ -189,7 +188,10 @@ errcode_t ext2fs_initialize(const char *name, int flags,
 		super->s_inode_size = EXT2_GOOD_OLD_INODE_SIZE;
 	}
 
-	set_field(s_checkinterval, EXT2_DFL_CHECKINTERVAL);
+	if (flags & EXT2_FLAG_MOUNT_FSCK_CHECK) {
+		set_field(s_max_mnt_count, EXT2_DFL_MAX_MNT_COUNT);
+		set_field(s_checkinterval, EXT2_DFL_CHECKINTERVAL);
+	}
 	super->s_mkfs_time = super->s_lastcheck = fs->now ? fs->now : time(NULL);
 
 	super->s_creator_os = CREATOR_OS;

misc/mke2fs.8.in +7 -0

index 2eead17..758307f 100644
@@ -18,6 +18,9 @@ mke2fs \- create an ext2/ext3/ext4 filesystem
 .I block-size
 ]
 [
+.B \-C
+]
+[
 .B \-f
 .I fragment-size
 ]

@@ -184,6 +187,10 @@ Check the device for bad blocks before creating the file system.  If
 this option is specified twice, then a slower read-write
 test is used instead of a fast read-only test.
 .TP
+.B \-C
+Set maximum mount count and interval between checks, so that a full fsck
+will be enforced when these counts are exceeded.
+.TP
 .BI \-E " extended-options"
 Set extended options for the filesystem.  Extended options are comma
 separated, and may take an argument using the equals ('=') sign.  The

misc/mke2fs.c +0 -0

index 0ba4a4c..fdeafe3 100644

misc/mke2fs.c +11 -5

 int	verbose;
 int	quiet;
 int	super_only;

@@ -1228,7 +1229,7 @@ profile_error:
 	}
 
 	while ((c = getopt (argc, argv,
-		    "b:cf:g:G:i:jl:m:no:qr:s:t:vE:FI:J:KL:M:N:O:R:ST:U:V")) != EOF) {
+		    "b:cf:g:G:i:jl:m:no:qr:s:t:vCE:FI:J:KL:M:N:O:R:ST:U:V")) != EOF) {
 		switch (c) {
 		case 'b':
 			blocksize = strtol(optarg, &tmp, 0);

@@ -1251,6 +1252,9 @@ profile_error:
 		case 'c':	/* Check for bad blocks */
 			cflag++;
 			break;
+		case 'C':	/* Set default check/fsck intervals */
+			Cflag++;
+			break;
 		case 'f':
 			size = strtoul(optarg, &tmp, 0);
 			if (size < EXT2_MIN_BLOCK_SIZE ||

@@ -2030,6 +2034,8 @@ int main (int argc, char *argv[])
 	 */
 	if (!quiet)
 		flags |= EXT2_FLAG_PRINT_PROGRESS;
+	if (Cflag)
+		flags |= EXT2_FLAG_MOUNT_FSCK_CHECK;
 	retval = ext2fs_initialize(device_name, flags, &fs_param, io_ptr, &fs);
 	if (retval) {
 		com_err(device_name, retval, _("while setting up superblock"));

@@ -2109,9 +2115,11 @@ int main (int argc, char *argv[])
 	 * don't check all the filesystems at the same time.  We use a
 	 * kludgy hack of using the UUID to derive a random jitter value.
 	 */
-	for (i = 0, val = 0 ; i < sizeof(fs->super->s_uuid); i++)
-		val += fs->super->s_uuid[i];
-	fs->super->s_max_mnt_count += val % EXT2_DFL_MAX_MNT_COUNT;
+	if (Cflag) {
+		for (i = 0, val = 0 ; i < sizeof(fs->super->s_uuid); i++)
+			val += fs->super->s_uuid[i];
+		fs->super->s_max_mnt_count += val % EXT2_DFL_MAX_MNT_COUNT;
+	}
 
 	/*
 	 * Override the creator OS, if applicable
 
--

To unsubscribe from this list: send the line "unsubscribe linux-ext4" in the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html

Quoted 10 lines · Eric Sandeen

The forced fsck often comes at unexpected and inopportune moments, and even enterprise customers are often caught by surprise when this happens.  Because a filesystem with an error condition will be marked as requiring fsck anyway, I submit that the time-based and mount-based checks are not particularly useful, and that administrators can schedule fscks on their own time, or tune2fs the enforced intervals if they so choose.  Patch #1 disables the intervals by default, and I've added a new mkfs option (-C) to turn on the old behavior of random, unexpected, time-consuming fscks at boot time.  ;)

I completely agree with you and I like the idea.

Quoted 71 lines

Signed-off-by Eric Sandeen <[email protected]>

---

lib/ext2fs/ext2fs.h +1 -0

index a204eb7..37584aa 100644
@@ -184,6 +184,7 @@ typedef struct ext2_file *ext2_file_t;
  #define EXT2_FLAG_64BITS             0x20000
  #define EXT2_FLAG_PRINT_PROGRESS     0x40000
  #define EXT2_FLAG_DIRECT_IO          0x80000
+#define EXT2_FLAG_MOUNT_FSCK_CHECK   0x100000
 
  /*
   * Special flag in the ext2 inode i_flag field that means that this is

lib/ext2fs/initialize.c +0 -0

index 7c38975..ff63a92 100644
@@ -155,7 +155,6 @@ errcode_t ext2fs_initialize(const char *name, int flags,

      set_field(s_log_block_size, 0); /* default blocksize: 1024 bytes */ set_field(s_log_frag_size, 0); /* default fragsize: 1024 bytes */ set_field(s_first_data_block, super->s_log_block_size ? 0 : 1);
-     set_field(s_max_mnt_count, EXT2_DFL_MAX_MNT_COUNT); set_field(s_errors, EXT2_ERRORS_DEFAULT); set_field(s_feature_compat, 0); set_field(s_feature_incompat, 0);

lib/ext2fs/initialize.c +4 -1

@@ -189,7 +188,10 @@ errcode_t ext2fs_initialize(const char *name, int flags,
               super->s_inode_size = EXT2_GOOD_OLD_INODE_SIZE;
       }
 
-     set_field(s_checkinterval, EXT2_DFL_CHECKINTERVAL);
+     if (flags & EXT2_FLAG_MOUNT_FSCK_CHECK) {
+             set_field(s_max_mnt_count, EXT2_DFL_MAX_MNT_COUNT);
+             set_field(s_checkinterval, EXT2_DFL_CHECKINTERVAL);
+     }
       super->s_mkfs_time = super->s_lastcheck = fs->now ? fs->now : time(NULL);
 
       super->s_creator_os = CREATOR_OS;

misc/mke2fs.8.in +0 -0

index 2eead17..758307f 100644
@@ -18,6 +18,9 @@ mke2fs \- create an ext2/ext3/ext4 filesystem

 .I block-size
 ]
 [
+.B \-C
+]
+[
 .B \-f
 .I fragment-size
 ]

misc/mke2fs.8.in +0 -0

@@ -184,6 +187,10 @@ Check the device for bad blocks before creating the file system.  If

 this option is specified twice, then a slower read-write test is used instead of a fast read-only test. .TP
+.B \-C

misc/mke2fs.8.in +3 -0

+Set maximum mount count and interval between checks, so that a full fsck
+will be enforced when these counts are exceeded.
+.TP

 .BI \-E " extended-options"
 Set extended options for the filesystem.  Extended options are comma separated, and may take an argument using the equals ('=') sign.  The

misc/mke2fs.c +0 -0

index 0ba4a4c..fdeafe3 100644

misc/mke2fs.c const char * device_name /* = NULL */; · line 83

Quoted 4 lines

misc/mke2fs.c +1 -0

@@ -80,6 +80,7 @@ const char * device_name /* = NULL */;
 
  /* Command line options */
  int  cflag;
+int  Cflag;

Uff, this is not really an issue, however would it be possible to have better name for Cflags :) ? it is just confusing.

Quoted 47 lines

misc/mke2fs.c +1 -1

  int  verbose;
  int  quiet;
  int  super_only;

@@ -1228,7 +1229,7 @@ profile_error:
       }
 
       while ((c = getopt (argc, argv,
-                 "b:cf:g:G:i:jl:m:no:qr:s:t:vE:FI:J:KL:M:N:O:R:ST:U:V")) != EOF) {
+                 "b:cf:g:G:i:jl:m:no:qr:s:t:vCE:FI:J:KL:M:N:O:R:ST:U:V")) != EOF) {
               switch (c) {

              case 'b':
                      blocksize = strtol(optarg, &tmp, 0);

misc/mke2fs.c +0 -0

@@ -1251,6 +1252,9 @@ profile_error:

              case 'c':       /* Check for bad blocks */ cflag++;
                      break;
+             case 'C':       /* Set default check/fsck intervals */
+                     Cflag++;
+                     break;
              case 'f':
                      size = strtoul(optarg, &tmp, 0);
                      if (size < EXT2_MIN_BLOCK_SIZE ||

misc/mke2fs.c +7 -3

@@ -2030,6 +2034,8 @@ int main (int argc, char *argv[])
        */
       if (!quiet)
               flags |= EXT2_FLAG_PRINT_PROGRESS;
+     if (Cflag)
+             flags |= EXT2_FLAG_MOUNT_FSCK_CHECK;
       retval = ext2fs_initialize(device_name, flags, &fs_param, io_ptr, &fs);
       if (retval) {
               com_err(device_name, retval, _("while setting up superblock"));

@@ -2109,9 +2115,11 @@ int main (int argc, char *argv[])
        * don't check all the filesystems at the same time.  We use a
        * kludgy hack of using the UUID to derive a random jitter value.
        */
-     for (i = 0, val = 0 ; i < sizeof(fs->super->s_uuid); i++)
-             val += fs->super->s_uuid[i];
-     fs->super->s_max_mnt_count += val % EXT2_DFL_MAX_MNT_COUNT;
+     if (Cflag) {
+             for (i = 0, val = 0 ; i < sizeof(fs->super->s_uuid); i++)
+                     val += fs->super->s_uuid[i];
+             fs->super->s_max_mnt_count += val % EXT2_DFL_MAX_MNT_COUNT;
+     }
 
       /*
        * Override the creator OS, if applicable

Quoted 5 lines

--
To unsubscribe from this list: send the line "unsubscribe linux-ext4" in the body of a message to [email protected]
More majordomo info at  http://vger.kernel.org/majordomo-info.html

--
--
To unsubscribe from this list: send the line "unsubscribe linux-ext4" in the body of a message to [email protected]
More majordomo info at  http://vger.kernel.org/majordomo-info.html

The forced fsck often comes at unexpected and inopportune moments, and even enterprise customers are often caught by surprise when this happens. Because a filesystem with an error condition will be marked as requiring fsck anyway, I submit that the time-based and mount-based checks are not particularly useful, and that administrators can schedule fscks on their own time, or tune2fs the enforced intervals if they so choose. Patch #1 disables the intervals by default, and I've added a new mkfs option (-C) to turn on the old behavior of random, unexpected, time-consuming fscks at boot time. ;)

Signed-off-by Eric Sandeen <[email protected]>

User namespace xattrs are generally useful, and I think extN is the only filesystem requiring a special mount option to enable them, when xattrs are otherwise available. So this change sets that mount option into the defaults.

Note that if xattrs are config'd off, this will lead to a mostly-harmless:

EXT4-fs (sdc1): (no)user_xattr options not supported

message at mount time...

Quoted 8 lines · Eric Sandeen

User namespace xattrs are generally useful, and I think extN is the only filesystem requiring a special mount option to enable them, when xattrs are otherwise available. So this change sets that mount option into the defaults.

Note that if xattrs are config'd off, this will lead to a mostly-harmless:

Quoted 3 lines

EXT4-fs (sdc1): (no)user_xattr options not supported

message at mount time...

Wouldn't it be more useful to change this in the kernel, instead of only changing it in the superblock for new filesystems?

Cheers, Andreas

Quoted 7 lines · Andreas Dilger

Quoted 5 lines · Eric Sandeen

User namespace xattrs are generally useful, and I think extN is the only filesystem requiring a special mount option to enable them, when xattrs are otherwise available. So this change sets that mount option into the defaults.

Quoted 2 lines

Wouldn't it be more useful to change this in the kernel, instead of only changing it in the superblock for new filesystems?

I thought about that; maybe it should happen in both places?

Quoted 11 lines · Eric Sandeen

Quoted 9 lines · Andreas Dilger

Quoted 5 lines · Eric Sandeen

User namespace xattrs are generally useful, and I think extN is the only filesystem requiring a special mount option to enable them, when xattrs are otherwise available. So this change sets that mount option into the defaults.

Wouldn't it be more useful to change this in the kernel, instead of only changing it in the superblock for new filesystems?

Quoted 1 line

I thought about that; maybe it should happen in both places?

Yes, agreed. Maybe we should do the same thing with posix ACL's as well? ACL's can only take away access as compared to the unix permission bits, so it's safe to enable ACL's.

- Ted

Whoops, I didn't realize that acls were in the same boat, of "config on and then turn on with a mount option"

Ok, sure, adding that makes sense to me too...

Is sticking it in the _initialize() function ok with you? It's a little odd, but ...

Quoted 5 lines · Eric Sandeen

Whoops, I didn't realize that acls were in the same boat, of "config on and then turn on with a mount option"

Ok, sure, adding that makes sense to me too...

Quoted 2 lines

Is sticking it in the _initialize() function ok with you? It's a little odd, but ...

I think it's probably better to put it in the misc/mke2fs.c; we set a number of superblock options there already. But either place is fine...

- Ted

Ok, if we do it in mke2fs.c that's probably better.

Or - I wonder how hard it'd be to make this into a mke2fs.conf option instead? I always forget about that route.

get_string_from_profile(... "default_mount_opts" ....) ?

User namespace xattrs are generally useful, and I think extN is the only filesystem requiring a special mount option to enable them, when xattrs are otherwise available. So this change sets that mount option into the defaults, via a mke2fs.conf option.

Note that if xattrs are config'd off, this will lead to a mostly-harmless:

EXT4-fs (sdc1): (no)user_xattr options not supported

message at mount time...

Signed-off-by Eric Sandeen <[email protected]>

Quoted 13 lines · Eric Sandeen

User namespace xattrs are generally useful, and I think extN is the only filesystem requiring a special mount option to enable them, when xattrs are otherwise available. So this change sets that mount option into the defaults, via a mke2fs.conf option.

Note that if xattrs are config'd off, this will lead to a mostly-harmless:

EXT4-fs (sdc1): (no)user_xattr options not supported

message at mount time...

Quoted 1 line

Signed-off-by Eric Sandeen <[email protected]>

You didn't run "make check" before submitting these patches. Please do so in the future! :-)

I'll fix up the expected test commit results for these two commits.

- Ted

Anaconda (the Fedora/RHEL installer) had been "fixing up" extN filesystems it created by setting the max mount count and check interval to 0, as well as adding user_xattr to filesystem mount options.

As part of their efforts to stop special-casing around upstream defaults, they've removed these changes upstream.

However, I'd like to at least propose that these changes be made default.

The forced fsck often comes at unexpected and inopportune moments, and even enterprise customers are often caught by surprise when this happens. Because a filesystem with an error condition will be marked as requiring fsck anyway,

I submit that the time-based and mount-based checks are not particularly useful, and that administrators can schedule fscks on their own time, or tune2fs the enforced intervals if they so choose.

Quoted 10 lines · Eric Sandeen

Anaconda (the Fedora/RHEL installer) had been "fixing up" extN filesystems it created by setting the max mount count and check interval to 0, as well as adding user_xattr to filesystem mount options.

As part of their efforts to stop special-casing around upstream defaults, they've removed these changes upstream.

However, I'd like to at least propose that these changes be made default.

I'd really prefer instead that the "lvcheck" script be included into the distro, instead of changing mke2fs. That achieves the same end result (periodic scrubbing of the filesystem to look for hidden errors), without introducing boot-time delays. Given the size of disks today and the undetected bit-error-rate (somewhere around 1/10^15 bits or 12TB), I think it is important that there be automated scrubbing of the filesystem.

I think the best place to put that script would be in the lvm tools (since it is applicable to multiple filesystems), which I think Eric has the most leverage in getting accepted (I've been but I'd be OK including it with e2fsprogs if there is pushback on that.

Quoted 4 lines

The forced fsck often comes at unexpected and inopportune moments, and even enterprise customers are often caught by surprise when this happens. Because a filesystem with an error condition will be marked as requiring fsck anyway,

Any decent RAID array does background scrubbing for integrity verification, it doesn't just wait until there is an uncorrectable error detected in the block device. If we can do something proactive to prevent this (i.e. lvcheck run by cron.weekly), it is worthwhile.

I think customers are equally surprised when their server fails (remount-ro/panic) due to the kernel detecting an error that might have been on disk for weeks or months.

Quoted 4 lines

I submit that the time-based and mount-based checks are not particularly useful, and that administrators can schedule fscks on their own time, or tune2fs the enforced intervals if they so choose.

I think you are projecting your own self-enlightenment onto users ;-). As we see on this list, there are many users that don't even back up their critical data, so IMHO taking out "safe by default" options is a step in the wrong direction.

Attached is my latest version of the lvcheck script, and a default /etc/lvcheck.conf script. It's been enhanced to include a usage message, command-line option parsing to override default parameters, and the ability to check snapshots of ext3/4 filesystems with an external journal.

Cheers, Andreas