[LU-14837] Layout corruption (lmm_oi) inside mdd_swap_layouts - Whamcloud Community JIRA

Details

Type: Bug
Resolution: Unresolved
Priority: Major
Fix Version/s: None
Affects Version/s: Lustre 2.10.0, Lustre 2.11.0, Lustre 2.12.0, Lustre 2.13.0
Labels:
None

Severity:
3
Rank (Obsolete):
9223372036854775807

Description

The following code in mdd_swap_layouts is not valid for PFL layout:

         /* set the file specific informations in lmm */
         if (fst_lmm != NULL) {
                 saved_oi = &info->mti_oa.o_oi;
                 mdd_get_lmm_oi(fst_lmm, saved_oi);
                 mdd_set_lmm_gen(fst_lmm, &snd_gen);
           ----> mdd_set_lmm_oi(fst_lmm, &snd_lmm->lmm_oi);
                 mdd_set_lmm_oi(snd_lmm, saved_oi);
         } else {

For PFL layout, snd_lmm does not point directly to a lov_mds_md structure but on a lov_comp_md_v1 structure.

We have to access to the different component entries defined inside lov_comp_md_v1 and then apply offsets to access to the lov_mds_md structures.

LFSK can correct this inconsistencies. this will increment the "inconsistencies_other/repaired_other" counter.

WARNING:

Do not try to run LFSCK on a 2.12.4 for this. This will corrupted every migrated layouts because of the ~~LU-13535~~ "Files truncated/corruption due to lfsck"

Attachments

Issue Links

duplicates

LU-13426 "lfs migrate" on DoM component clobbers LOV EA FID

Resolved

is related to

LU-13535 Files truncated/corruption due to lfsck

Resolved

Activity

[LU-14837] Layout corruption (lmm_oi) inside mdd_swap_layouts

Etienne Aujames made changes - 21/Jul/21 11:33 AM

Affects Version/s		New: Lustre 2.10.0 [ 12204 ]
Affects Version/s	Original: Lustre 2.14.0 [ 14490 ]

Etienne Aujames made changes - 21/Jul/21 11:32 AM

Affects Version/s		New: Lustre 2.11.0 [ 13091 ]
Affects Version/s	Original: Lustre 2.10.0 [ 12204 ]

Etienne Aujames made changes - 21/Jul/21 11:31 AM

Affects Version/s

New: Lustre 2.14.0 [ 14490 ]

Etienne Aujames made changes - 09/Jul/21 6:20 PM

Description

Original: The following code in mdd_swap_layouts is not valid for PFL layout:
{code:java}
         /* set the file specific informations in lmm */
         if (fst_lmm != NULL) {
                 saved_oi = &info->mti_oa.o_oi;
                 mdd_get_lmm_oi(fst_lmm, saved_oi);
                 mdd_set_lmm_gen(fst_lmm, &snd_gen);
           ----> mdd_set_lmm_oi(fst_lmm, &snd_lmm->lmm_oi);
                 mdd_set_lmm_oi(snd_lmm, saved_oi);
         } else {

{code}
For PFL layout, snd_lmm does not point directly to a lov_mds_md structure but on a lov_comp_md_v1 structure.

We have to access to the different component entries define inside lov_comp_md_v1 and then apply offsets to access to the lov_mds_md structures.

LFSK can correct this inconsistencies. this will increment the "inconsistencies_other/repaired_other" counter.

{color:#de350b}*WARNING*{color}:

Do not try to run LFSCK on a 2.12.4 for this. This will corrupted *every migrated layouts* because of the ~~LU-13535~~ "Files truncated/corruption due to lfsck"

New: The following code in mdd_swap_layouts is not valid for PFL layout:
{code:java}
         /* set the file specific informations in lmm */
         if (fst_lmm != NULL) {
                 saved_oi = &info->mti_oa.o_oi;
                 mdd_get_lmm_oi(fst_lmm, saved_oi);
                 mdd_set_lmm_gen(fst_lmm, &snd_gen);
           ----> mdd_set_lmm_oi(fst_lmm, &snd_lmm->lmm_oi);
                 mdd_set_lmm_oi(snd_lmm, saved_oi);
         } else {

{code}
For PFL layout, snd_lmm does not point directly to a lov_mds_md structure but on a lov_comp_md_v1 structure.

We have to access to the different component entries defined inside lov_comp_md_v1 and then apply offsets to access to the lov_mds_md structures.

LFSK can correct this inconsistencies. this will increment the "inconsistencies_other/repaired_other" counter.

{color:#de350b}*WARNING*{color}:

Do not try to run LFSCK on a 2.12.4 for this. This will corrupted *every migrated layouts* because of the ~~LU-13535~~ "Files truncated/corruption due to lfsck"

Etienne Aujames made changes - 09/Jul/21 5:51 PM

Link

New: This issue duplicates ~~LU-13426~~ [ ~~LU-13426~~ ]

Etienne Aujames made changes - 09/Jul/21 5:49 PM

Affects Version/s

Original: Lustre 2.14.0 [ 14490 ]

Etienne Aujames made changes - 09/Jul/21 5:43 PM

Link

New: This issue is related to CEA-5 [ CEA-5 ]

Etienne Aujames made changes - 09/Jul/21 5:43 PM

Link

New: This issue is related to ~~LU-13535~~ [ ~~LU-13535~~ ]

Etienne Aujames made changes - 09/Jul/21 5:39 PM

Description

Original: The following code in mdd_swap_layouts is not valid for PFL layout:
{code:java}
         /* set the file specific informations in lmm */
         if (fst_lmm != NULL) {
                 saved_oi = &info->mti_oa.o_oi;
                 mdd_get_lmm_oi(fst_lmm, saved_oi);
                 mdd_set_lmm_gen(fst_lmm, &snd_gen);
           ----> mdd_set_lmm_oi(fst_lmm, &snd_lmm->lmm_oi);
                 mdd_set_lmm_oi(snd_lmm, saved_oi);
         } else {

{code}
For PFL layout, snd_lmm does not point directly to a lov_mds_md structure but on a lov_comp_md_v1 structure.

We have to access to the different component entries define inside lov_comp_md_v1 and then apply offsets to access to the lov_mds_md structures.

LFSK can correct this inconsistencies. this will increment the "inconsistencies_other/repaired_other" counter.

WARNING:

Do not try to run LFSCK on a 2.12.4 for this. This will corrupted every migrated layouts because of the

New: The following code in mdd_swap_layouts is not valid for PFL layout:
{code:java}
         /* set the file specific informations in lmm */
         if (fst_lmm != NULL) {
                 saved_oi = &info->mti_oa.o_oi;
                 mdd_get_lmm_oi(fst_lmm, saved_oi);
                 mdd_set_lmm_gen(fst_lmm, &snd_gen);
           ----> mdd_set_lmm_oi(fst_lmm, &snd_lmm->lmm_oi);
                 mdd_set_lmm_oi(snd_lmm, saved_oi);
         } else {

{code}
For PFL layout, snd_lmm does not point directly to a lov_mds_md structure but on a lov_comp_md_v1 structure.

We have to access to the different component entries define inside lov_comp_md_v1 and then apply offsets to access to the lov_mds_md structures.

LFSK can correct this inconsistencies. this will increment the "inconsistencies_other/repaired_other" counter.

{color:#de350b}*WARNING*{color}:

Do not try to run LFSCK on a 2.12.4 for this. This will corrupted *every migrated layouts* because of the ~~LU-13535~~ "Files truncated/corruption due to lfsck"

Etienne Aujames made changes - 09/Jul/21 5:29 PM

Description

New: The following code in mdd_swap_layouts is not valid for PFL layout:
{code:java}
         /* set the file specific informations in lmm */
         if (fst_lmm != NULL) {
                 saved_oi = &info->mti_oa.o_oi;
                 mdd_get_lmm_oi(fst_lmm, saved_oi);
                 mdd_set_lmm_gen(fst_lmm, &snd_gen);
           ----> mdd_set_lmm_oi(fst_lmm, &snd_lmm->lmm_oi);
                 mdd_set_lmm_oi(snd_lmm, saved_oi);
         } else {

{code}
For PFL layout, snd_lmm does not point directly to a lov_mds_md structure but on a lov_comp_md_v1 structure.

We have to access to the different component entries define inside lov_comp_md_v1 and then apply offsets to access to the lov_mds_md structures.

LFSK can correct this inconsistencies. this will increment the "inconsistencies_other/repaired_other" counter.

WARNING:

Do not try to run LFSCK on a 2.12.4 for this. This will corrupted every migrated layouts because of the

People

Assignee:: Etienne Aujames

Reporter:: Etienne Aujames

Votes:: 0 Vote for this issue

Watchers:: 3 Start watching this issue

Dates

Created:: 09/Jul/21 5:00 PM

Updated:: 21/Jul/21 11:33 AM