installation fail according lustre_server_rocky8.10 #1
Comments
|
I'm not too sure what can be done from afar to offer help (I might need to test if Rocky changed at some point...), but there are some pointers. If you check your output, you see the following; Basically, application of the patches failed. There are at least two scenarios I can imagine:
Maybe write a step by step log of what you do to get to where you are? |
|
hi, expert , i show the error messages
could you help me again .or do you any error message. i will show to you
cd /root/lustre-release/ldiskfs/linux-stage/fs/ext4
[root@Lustre ext4]# cat xattr.c.rej
--- fs/ext4/xattr.c
+++ fs/ext4/xattr.c
@@ -1626,6 +1656,7 @@ static int ext4_xattr_set_entry(struct e
int in_inode = i->in_inode;
struct inode *old_ea_inode = NULL;
struct inode *new_ea_inode = NULL;
+ struct delayed_iput_work *diwork = NULL;
size_t old_size, new_size;
int ret;
@@ -1859,7 +1894,7 @@ update_hash:
ret = 0;
out:
- iput(old_ea_inode);
+ delayed_iput(old_ea_inode, diwork);
iput(new_ea_inode);
return ret;
}
do you have any ideas ?
夏天
***@***.***
Original
From: "Detlev C M" ***@***.***>
Sent Time: 2024-11-21- 16:22
To: "DetlevCM/lustre-build-instructions" ***@***.***>
Cc: "xia-MM" ***@***.***> , "Author" ***@***.***>
Subject: Re: [DetlevCM/lustre-build-instructions] installation fail accordinglustre_server_rocky8.10 (Issue #1)
I'm not too sure what can be done from afar to offer help (I might need to test if Rocky changed at some point...), but there are some pointers.
If you check your output, you see the following;
... rhel8.7/ext4-introduce-EXT4_BG_TRIMMED-to-optimize-fstrim.patch rhel8/ext4-ialloc-uid-gid-and-pass-owner-down.patch base/ext4-projid-xattrs.patch rhel8.5/ext4-enc-flag.patch rhel8/ext4-ext-merge.patch base/ext4-delayed-iput.patch2 out of 4 hunks FAILED -- saving rejects to file fs/ext4/xattr.c.rej
Basically, application of the patches failed.
There are at least two scenarios I can imagine:
When I first set up a lustre VM as a test, I found that the ext4 code was not compatible with the patch files and I had to tweak the patch file. Then the kernel updated, all worked... It may be necessary again.
Please check the error in detail, specifically the file that triggers it.
Did you try applying the patches more than once? If yes, this will create a mess. In this case, I would recommend unpacking the kernel again and copying over files as appropriate, followed by patching.
Maybe write a step by step log of what you do to get to where you are?
—
Reply to this email directly, view it on GitHub, or unsubscribe.
You are receiving this because you authored the thread.
|
|
hi, sir: i solve it . when i re-edit the ext4-delayed-iput.patch file ,then delete the following line : Index: b2_15_linux-4.18.0-425.3.1.el8/fs/ext4/xattr.c--- b2_15_linux-4.18.0-425.3.1.el8.orig/fs/ext4/xattr.c +struct delayed_iput_work {
+};
+}
+}
@@ -1672,7 +1703,11 @@ static int ext4_xattr_set_entry(struct e
@@ -1825,7 +1860,7 @@ update_hash: out:
} then execute the under command |
|
It looks like the formatting broke a bit... Good to hear that you found the solution. Honestly, it is a bit concerning that the build procedure does not seem to be entirely stable and had issues from time to time... |
|
hi,expert during the command configure lustre, can i choice ignore to install the ext4 patch file ? |
I'm just a user of the code in the end 😅 As to the patch: I do not think it can be ignored, because lustre uses a modified ext4 file system. (Not sure if it can also use ZFS for the OSTs.) May I suggest the lustre mailing list? - It would be a better place for questions that really address the inner workings of lustre: https://www.lustre.org/mailing-lists/ |
|
thx |
|
sir, want to require another problem: |
|
@xia-MM Just lie this, no, sorry... Are you building this on Redhat or Rocky? - I recently ran through the rocky install for a test system and 8.10 worked without issues (except for a slightly different build number in the kernel). When googling the error message, one finds that the same issue has occurred in the past with Lustre 2.9.x ... - I'm not sure sure what the preferred solution is. |
|
hi, sir, i build the lustre on Rocky 8.10. which kernel version do you install ? this is my kernel version : kernel-4.18.0-553.32.1.el8_10 |
Checking the test system (VM actually), this is the kernel used: And as mentioned before, I only had to tweak the odd build number here or there. |
|
hi, sir i want to ask to you private question, could you give me your e-mail ? i want to require to you some problem when i meet some difficult things during learn luster . i like to learn lustre follow you. |
|
hi, sir, did you have ever install lustre+ZFS on the ubuntu ? i try to compile zfs +lustre, when i try to the command " ./configure --with-linux=/root/linux-source-5.4.0/ --enable-server --enable-modules --with-zfs=/root/zfs/zfs-2.1.2/ --enable-ldiskfs " |
Short answer, no, I never tried. |
|
hi, sir, did you have install packmaker for MGS/MDS? how to install agent : ocf:llnl:lustre ? |
Sorry, no. I can only point you at the official lustre documentation: https://wiki.whamcloud.com/display/PUB/Using+Pacemaker+1.1+with+a+Lustre+File+System |
|
hi,sir, which high available software do you choice ? when you build a lustre environment for MGS/MDS/OSS ?
|
|
hi,sir,, have your ever meet the following error during build zfs-osd . the lustre version:2.15.5 . zfs version: 2.11. LD [M] /root/lustre-release/lustre/osc/osc.o |
|
I haven't personally built lustre with zfs, so I cannot comment. |
|
hi,sir, which high available software do you choice ? when you build a lustre environment for MGS/MDS/OSS ? |
As to patches: when patches are release, to the best of my knowledge, these are applied to the source code you would need to recompile the software and replace binaries/libraries. |
|
hi, sorry for disturb you again. did you meeting the following error messages ,when install e2fsprogs software on the ubuntu.
|
|
I think the original issue here is the same as for the issue I opened right now for the 9.4 instructions. The problem is that the instructions reference an older version of the 9.4 or 8.10 kernel src.rpm In the case of 9.4 this is the ldisk patch that no longer applies and causes the build to break ./ldiskfs/kernel_patches/patches/rhel9.4/ext4-delayed-iput.patch Same error: 2 out of 4 hunks FAILED -- saving rejects to file fs/ext4/xattr.c.rej Here as well you will need to update the instructions to point to the most recent 8.10 kernel src rpm. |
|
hi, RonnieSahlberg: OK, i will be try .and i will be reply to the result. |
Indeed, there is a problem with the kernel moving: I recently encountered an issue with Rocky Linux where installing 9.4 now will lead to up to date sources for a 9.5 kernel being downloaded... - The rpms no longer build - until the patches are available (and I hope they are now) as something changed in the kernel. |
|
always up to date lustre's version , so that can get new feature .but will be meet new bug or new problem . everytimes when i deploy produce environment at the customer field. i strong aggainst recomment lastest version. |
This is where the prebuilt binaries from Whamcloud come into play. As it stands, I don't think there is a good short term answer due to the number of patches and adjustments made on the lustre side. |
|
The lustre developer do not have any influence of what changes RedHat makes between minor releases which can cause the Lustre out-of-tree patches to break. Maybe you need to change the instructions to lock the lustre version to a specific sha1 commit that is known to work. |
|
hi, my meaning, if i am according ChangeLogs strictly about kernel version and lustre version . compile with source code, maybe i will be unsuccessful . |
|
What i means is, If you want to use kernel kernel-5.14.0-427.37.1.el9_4 and build lustre-release But IF you want it to work, you can checkout an older version of lustre-release before the incompatible changes for post 5.14.0-37.1 was made. do this: and that will allow you to build against a 5.14.0-427.37.1 kernel. |
|
hi,sir, intend to consult another question: when i try to installation Rocky Linux9.4 , display the following messages: error: unable to find a match: kernel-abi-whitelists . how to get kernel-abi-whitelists ? |
|
The KABI lists are in the SOURCES directory when you unpack the srpm. it is the kabi files. Maybe you can disable the kabi checks? The kabi checks are just for redhat to make sure that they do not change any of the APIs that important out of tree modules (that redhat cares about) change. It s a list of ~700 out of ~30.000 kernel symbols that are important that they do not change because some important third party vendor modules depend on. |
hi, expert,
i am a interest lustre , this is my first time to installation lustre . try to many times according your document , meeting the following error messages. how to fix it ,could you help me .
if test -n "" ; then
cp linux-stage/include/linux;
fi
if test -n "" ; then
cp linux-stage/include/uapi/linux;
fi
if test -n "/root/kernel/rpmbuild/BUILD/kernel-4.18.0-553.27.1.el8_10/linux-4.18.0-553.27.1.el8_lustre.x86_64/include/trace/events/ext4.h" ; then
cp /root/kernel/rpmbuild/BUILD/kernel-4.18.0-553.27.1.el8_10/linux-4.18.0-553.27.1.el8_lustre.x86_64/include/trace/events/ext4.h linux-stage/include/trace/events;
fi
Applying ext4 patches: rhel8.1/ext4-inode-version.patch suse15/ext4-lookup-dotdot.patch suse15/ext4-print-inum-in-htree-warning.patch rhel8/ext4-prealloc.patch ubuntu18/ext4-osd-iop-common.patch rhel8.7/ext4-misc.patch rhel8.7/ext4-mballoc-extra-checks.patch rhel8.7/ext4-hash-indexed-dir-dotdot-update.patch rhel8.1/ext4-kill-dx-root.patch rhel8.7/ext4-mballoc-pa-free-mismatch.patch rhel8.4/ext4-data-in-dirent.patch rhel8/ext4-nocmtime.patch base/ext4-htree-lock.patch rhel8.7/ext4-pdirop.patch rhel8/ext4-deep-tree.patch rhel8/ext4-max-dir-size.patch rhel8.7/ext4-corrupted-inode-block-bitmaps-handling-patches.patch ubuntu18/ext4-give-warning-with-dir-htree-growing.patch ubuntu18/ext4-jcb-optimization.patch rhel8.2/ext4-attach-jinode-in-writepages.patch rhel8/ext4-dont-check-before-replay.patch rhel7.6/ext4-use-GFP_NOFS-in-ext4_inode_attach_jinode.patch rhel7.6/ext4-export-orphan-add.patch rhel8/ext4-export-mb-stream-allocator-variables.patch rhel8/ext4-simple-blockalloc.patch rhel8.7/ext4-mballoc-skip-uninit-groups-cr0.patch rhel8.7/ext4-mballoc-prefetch.patch rhel8.3/ext4-xattr-disable-credits-check.patch base/ext4-no-max-dir-size-limit-for-iam-objects.patch rhel7.6/ext4-dquot-commit-speedup.patch rhel8.7/ext4-introduce-EXT4_BG_TRIMMED-to-optimize-fstrim.patch rhel8/ext4-ialloc-uid-gid-and-pass-owner-down.patch base/ext4-projid-xattrs.patch rhel8.5/ext4-enc-flag.patch rhel8/ext4-ext-merge.patch base/ext4-delayed-iput.patch2 out of 4 hunks FAILED -- saving rejects to file fs/ext4/xattr.c.rej
make[2]: *** [autoMakefile:679: sources] Error 1
make[2]: Leaving directory '/root/lustre-release/ldiskfs'
make[1]: *** [autoMakefile:731: all-recursive] Error 1
make[1]: Leaving directory '/root/lustre-release'
make: *** [autoMakefile:585: all] Error 2
The text was updated successfully, but these errors were encountered: