Bug 1783946
| Summary: | The volgroup option --reserved-percent doesn't work | ||||||
|---|---|---|---|---|---|---|---|
| Product: | Red Hat Enterprise Linux 8 | Reporter: | Qin Yuan <qiyuan> | ||||
| Component: | python-blivet | Assignee: | Blivet Maintenance Team <blivet-maint-list> | ||||
| Status: | CLOSED ERRATA | QA Contact: | Release Test Team <release-test-team-automation> | ||||
| Severity: | high | Docs Contact: | |||||
| Priority: | high | ||||||
| Version: | 8.2 | CC: | blivet-maint-list, cshao, dlehman, jrusz, jstodola, lmiksik, lsvaty, mavital, mtessun, nlevy, peyu, pjanda, qiyuan, rhandlin, sbonazzo, shlei, vtrefny, weiwang, yaniwang, yturgema | ||||
| Target Milestone: | rc | Keywords: | Regression, TestCaseNeeded, ZStream | ||||
| Target Release: | 8.3 | ||||||
| Hardware: | Unspecified | ||||||
| OS: | Unspecified | ||||||
| Whiteboard: | |||||||
| Fixed In Version: | python-blivet-3.2.2-1.el8 | Doc Type: | If docs needed, set a value | ||||
| Doc Text: | Story Points: | --- | |||||
| Clone Of: | |||||||
| : | 1841131 (view as bug list) | Environment: | |||||
| Last Closed: | 2020-11-04 03:22:23 UTC | Type: | Bug | ||||
| Regression: | --- | Mount Type: | --- | ||||
| Documentation: | --- | CRM: | |||||
| Verified Versions: | Category: | --- | |||||
| oVirt Team: | --- | RHEL 7.3 requirements from Atomic Host: | |||||
| Cloudforms Team: | --- | Target Upstream Version: | |||||
| Embargoed: | |||||||
| Bug Depends On: | |||||||
| Bug Blocks: | 1841131 | ||||||
| Attachments: |
|
||||||
|
Description
Qin Yuan
2019-12-16 10:07:53 UTC
Has this ever worked ? Yes, it has been working ever since Bug 1131247 was fixed. The update of Bug 1131247 brings the following changes: * If you create a LVM thin pool with automatic partitioning, 20 % of the volume group size is reserved, with a minimum of 1 GiB and a maximum of 100 GiB. * If you use the `logvol --thinpool --grow` command in a Kickstart file, the thin pool will grow to the maximum possible size, which means no space will be left for it in the volume group to grow. In this case, you can use the `volgroup --reserved-space` or `volgroup --reserved-percent` command to leave some space in the volume group reserved, which is recommended. Right, I checked on both also, anaconda reserves some space (not really 10%) on 7.7, but it doesn't work the same on 8. Looks like a blivet issue, David any idea ? I added a link to an untested pull request. If you tell me a tree/compose or blivet version I can make an updates image to allow you to test the change. The latest RHVH 4.4 is consuming python-blivet-3.1.0-20.el8 To test the proposed patch, please add the following to the installer boot/kernel command line: inst.updates=http://people.redhat.com/~dlehman/updates/lvm-thin-reserved.0.img Please let me know how it goes and attach logs if there is a failure, whether it be the same one or something new. Thanks. Tested RHVH-UNSIGNED-ISO-4.4-RHEL-8-20200318.0-RHVH-x86_64-dvd1.iso with the given inst.updates, --reserved-percent still doesn't work. Created attachment 1671918 [details]
installation logs in /tmp
Due to comment#2 adding Regression keyword, and targeting to 4.4.0. Reproduced on RHEL-8.2 with the following ks: text keyboard --vckeymap=us --xlayouts='us' lang en_US.UTF-8 rootpw redhat timezone Europe/Prague --isUtc reboot # Disk partitioning information clearpart --all --initlabel bootloader --location=mbr reqpart --add-boot part pv.01 --size=8000 volgroup vg pv.01 --reserved-percent=10 logvol swap --fstype=swap --name=swap --vgname=vg --size=1000 logvol none --name=pool --vgname=vg --thinpool --size=1 --grow logvol / --name=root --vgname=vg --thin --poolname=pool --size=4000 --grow %packages @base %end [root@localhost ~]# vgs VG #PV #LV #SN Attr VSize VFree vg 1 3 0 wz--n- <7.81g 0 [root@localhost ~]# Qin Yuan can you please test this with this new updates image: http://file.emea.redhat.com/~vtrefny/img/rhbz1783946.img It contains fix previously mentioned by Dave Lehman and also fix for rhbz#1737490 which is related to this issue. I tested this updates image with kickstart posted by Jan Stodola in comment #15 and it worked for me (created VG had 800 MB free space which is 10 % as requested in the kickstart). Tested the update img: 1. Failed to install on dirty disks, same as https://bugzilla.redhat.com/show_bug.cgi?id=1766498 2. Install on clean disk, use kickstart in comment #0, VFree is 9% of VSize, not 10%. In vgdisplay, Free size is almost 10% of Alloc size, see: [root@ati-local-02 ~]# vgs --units m VG #PV #LV #SN Attr VSize VFree rhvh 1 11 0 wz--n- 199996.00m 18120.00m [root@ati-local-02 ~]# vgdisplay --units m --- Volume group --- VG Name rhvh System ID Format lvm2 Metadata Areas 1 Metadata Sequence No 40 VG Access read/write VG Status resizable MAX LV 0 Cur LV 11 Open LV 8 Max PV 0 Cur PV 1 Act PV 1 VG Size 199996.00 MiB PE Size 4.00 MiB Total PE 49999 Alloc PE / Size 45469 / 181876.00 MiB Free PE / Size 4530 / 18120.00 MiB VG UUID EZKzwq-X6V3-03By-lTMm-UFw6-ivpZ-GVqC6j Shouldn't VFree=VSize*10% ? I've tested this again with kickstart from the comment #0 and it still works for me, I have 5000 of 49999 physical extents free: # vgdisplay rhvh --units=m --- Volume group --- VG Name rhvh System ID Format lvm2 Metadata Areas 1 Metadata Sequence No 9 VG Access read/write VG Status resizable MAX LV 0 Cur LV 4 Open LV 3 Max PV 0 Cur PV 1 Act PV 1 VG Size 199996.00 MiB PE Size 4.00 MiB Total PE 49999 Alloc PE / Size 44999 / 179996.00 MiB Free PE / Size 5000 / 20000.00 MiB VG UUID w5nA38-PCZQ-TOFh-M6HC-Wpfp-8h0L-7DQom3 Your vgdisplay output shows 11 total LVs, but the kickstart contains only 4 (swap, thinpool and 2 thin LVs for / and /var) -- this could be reason behind the difference in our tests. I used RHVH-4.4-20200507.1-RHVH-x86_64-dvd1.iso + inst.updates, rhvh will create required NIST LVs automatically if they are not specified in ks, and also rhvh layer, see: [root@ati-local-02 ~]# lvs LV VG Attr LSize Pool Origin Data% Meta% Move Log Cpy%Sync Convert home rhvh Vwi-aotz-- 1.00g pool 4.79 pool rhvh twi-aotz-- 159.89g 5.12 2.14 rhvh-4.4.0.18-0.20200507.0 rhvh Vwi---tz-k 145.24g pool root rhvh-4.4.0.18-0.20200507.0+1 rhvh Vwi-aotz-- 145.24g pool rhvh-4.4.0.18-0.20200507.0 4.74 root rhvh Vri---tz-k 145.24g pool swap rhvh -wi-ao---- 15.72g tmp rhvh Vwi-aotz-- 1.00g pool 4.82 var rhvh Vwi-aotz-- <14.65g pool 3.64 var_crash rhvh Vwi-aotz-- 10.00g pool 2.24 var_log rhvh Vwi-aotz-- 8.00g pool 2.45 var_log_audit rhvh Vwi-aotz-- 2.00g pool 4.77 Even if I define all the NIST LVs in ks, like below, vgs shows the same result: zerombr clearpart --all bootloader --location=mbr reqpart --add-boot part pv.01 --size=200000 volgroup rhvh pv.01 --reserved-percent=10 logvol swap --fstype=swap --name=swap --vgname=rhvh --recommended logvol none --name=pool --vgname=rhvh --thinpool --size=1 --grow logvol / --fstype=ext4 --name=root --vgname=rhvh --thin --poolname=pool --size=10000 --grow logvol /var --fstype=ext4 --name=var --vgname=rhvh --thin --poolname=pool --size=15000 logvol /var/log --fstype=ext4 --name=var_log --vgname=rhvh --thin --poolname=pool --size=8192 logvol /var/log/audit --fstype=ext4 --name=var_log_audit --vgname=rhvh --thin --poolname=pool --size=2048 logvol /home --fstype=ext4 --name=home --vgname=rhvh --thin --poolname=pool --size=1024 logvol /tmp --fstype=ext4 --name=tmp --vgname=rhvh --thin --poolname=pool --size=1024 logvol /var/crash --fstype=ext4 --name=var_crash --vgname=rhvh --thin --poolname=pool --size=10240 I unfortunately wasn't able to find the RHVH-4.4-20200507.1-RHVH-x86_64-dvd1.iso image on our devel mirrors so I couldn't test this but creating additional LVs and snapshots in the %post scripts will change the free space in the VG -- the 1% decrease in the free space is most likely caused by the pool metadata and/or pmspare grow. The 1% decrease in free space is caused by pool_tmeta and pmspare. Using ks in comment #22, before runing %post section, check lv, vg: [anaconda root@dell-per730-34 ~]# lvs -a LV VG Attr LSize Pool Origin Data% Meta% Move Log Cpy%Sync Convert home rhvh Vwi-aotz-- 1.00g pool 4.79 [lvol0_pmspare] rhvh ewi------- 84.00m pool rhvh twi-aotz-- 159.89g 4.50 15.10 [pool_tdata] rhvh Twi-ao---- 159.89g [pool_tmeta] rhvh ewi-ao---- 84.00m root rhvh Vwi-aotz-- 123.24g pool 4.90 swap rhvh -wi-ao---- 15.72g tmp rhvh Vwi-aotz-- 1.00g pool 4.79 var rhvh Vwi-aotz-- <14.65g pool 3.79 var_crash rhvh Vwi-aotz-- 10.00g pool 2.24 var_log rhvh Vwi-aotz-- 8.00g pool 2.45 var_log_audit rhvh Vwi-aotz-- 2.00g pool 4.76 [anaconda root@dell-per730-34 ~]# vgdisplay --units m --- Volume group --- VG Name rhvh System ID Format lvm2 Metadata Areas 1 Metadata Sequence No 19 VG Access read/write VG Status resizable MAX LV 0 Cur LV 9 Open LV 8 Max PV 0 Cur PV 1 Act PV 1 VG Size 199996.00 MiB PE Size 4.00 MiB Total PE 49999 Alloc PE / Size 44999 / 179996.00 MiB Free PE / Size 5000 / 20000.00 MiB VG UUID n37jk9-6u85-IwJZ-0L7g-V3HF-g6DY-YoOCyO As you can see pmspare and pool_tmeta are both 84m, vgfree is 10% of vgsize. After %post section finished, check lv, vg: [root@ati-local-02 ~]# lvs -a LV VG Attr LSize Pool Origin Data% Meta% Move Log Cpy%Sync Convert home rhvh Vwi-aotz-- 1.00g pool 4.79 [lvol0_pmspare] rhvh ewi------- 1.00g pool rhvh twi-aotz-- 159.89g 4.58 2.09 [pool_tdata] rhvh Twi-ao---- 159.89g [pool_tmeta] rhvh ewi-ao---- 1.00g rhvh-4.4.0.18-0.20200507.0 rhvh Vwi---tz-k 123.24g pool root rhvh-4.4.0.18-0.20200507.0+1 rhvh Vwi-aotz-- 123.24g pool rhvh-4.4.0.18-0.20200507.0 4.90 root rhvh Vri---tz-k 123.24g pool swap rhvh -wi-ao---- 15.72g tmp rhvh Vwi-aotz-- 1.00g pool 4.80 var rhvh Vwi-aotz-- <14.65g pool 3.63 var_crash rhvh Vwi-aotz-- 10.00g pool 2.24 var_log rhvh Vwi-aotz-- 8.00g pool 2.50 var_log_audit rhvh Vwi-aotz-- 2.00g pool 4.77 [root@ati-local-02 ~]# vgdisplay --units m --- Volume group --- VG Name rhvh System ID Format lvm2 Metadata Areas 1 Metadata Sequence No 35 VG Access read/write VG Status resizable MAX LV 0 Cur LV 11 Open LV 8 Max PV 0 Cur PV 1 Act PV 1 VG Size 199996.00 MiB PE Size 4.00 MiB Total PE 49999 Alloc PE / Size 45469 / 181876.00 MiB Free PE / Size 4530 / 18120.00 MiB VG UUID n37jk9-6u85-IwJZ-0L7g-V3HF-g6DY-YoOCyO pool_tmeta and pmspare are both extended to 1024m, it requires (1024m-84m)*2=1880m from free space of vg. And vgfree is reduced from 20000m to 18120m. In rhvh post section, there is a flow to increase pool meta to 1G. Now we can say the reserve-percent option works as expected. Test for this is created and merged. Verified on RHEL-8.3.0-20200616.0 x86_64 The fixed version "python-blivet-3.2.2-1.el8" is for 8.3, so the target milestone should be changed to 8.3 but not 8.2. Since the problem described in this bug report should be resolved in a recent advisory, it has been closed with a resolution of ERRATA. For information on the advisory (python-blivet bug fix and enhancement update), and where to find the updated files, follow the link below. If the solution does not work for you, open a new bug report. https://access.redhat.com/errata/RHBA-2020:4728 |