©XSIBackup-Free: Free Backup Software for ©VMWare ©ESXi

Forum ©XSIBackup: ©VMWare ©ESXi Backup Software


You are not logged in.

#1 2019-02-26 13:29:01

m_vk
Member
Registered: 2019-02-20
Posts: 5

XSITools error "sh: Bad block detected: 52428800"

This has been bothering us since we started using XSIBackup-PRO:

When using --backup-prog=xsitools:z, backup job (run from console) ends with:

...
---------------------------------------------------------------------------------------------------------------------------
Please wait while the (c)XSITools repository hash tables are fast checked...
---------------------------------------------------------------------------------------------------------------------------
sh: Bad block detected: 52428800
1: bad number
[VM0] error: chunk hashes fast check failed, detail:
---------------------------------------------------------------------------------------------------------------------------
The ESXi configuration was saved to "/vmfs/volumes/backup/xsitools-repos/20190200000000/20190226075817"
---------------------------------------------------------------------------------------------------------------------------
Errors detected in backup, check logs
...

What exactly does it want to say? Does it mean disk block, or that 50 MB chunk xsitools uses? Anyway 52428800 doesn't seem to refer to block ID, as it is exatly the xsitools chunk size (50 x 1024 x 1024).

Also no detail is given.

Subsequent repository check

--check-xsitoolsrepo="/vmfs/volumes/backup/xsitools-repos/20190200000000" --del-badblocks=yes

will find and delete 2 bad blocks, but next backup reports the same error!

ESX hardware is brand new Dell PowerEdge server with enterprise class SSDs. Backup target is NFS storage - initially Synology NAS, now dedicated Linux server, but errors keep showing on both of them (actually we switched to custom Linux machine after we me those "bad block detected" errors in backup to Synology).

ESX is

VMware ESXi 6.7.0 build-10302608

Thanks for reply.

Last edited by m_vk (2019-02-26 13:30:23)

Offline

#2 2019-02-26 13:32:10

m_vk
Member
Registered: 2019-02-20
Posts: 5

Re: XSITools error "sh: Bad block detected: 52428800"

Addendum: we are using "monthly" repository rollover, e.g.

--backup-point="/vmfs/volumes/backup/xsitools-repos/$(date +%Y%m'00000000')"

but this is probably apparent from previous output.

Offline

#3 2019-02-26 19:33:26

admin
Administrator
Registered: 2017-04-21
Posts: 2,055

Re: XSITools error "sh: Bad block detected: 52428800"

We would need more info. The best you can do is to avoid the fast check, which actually just compares sizes and perform a full check on the repo. That will take longer, but will throw more information.

--certify-backup=yes, instead of --certify-backup=fast

The text 1: bad number is a shell error, some amount is being compared to something that's not a number, but that in turn comes from somewhere else, so the full log is neccessary to offer more advice. Please, contact support to provide the details of your case.

A bad block issue is something very rare and is usually related to hardware flaws. In your case, it seems to be a secondary error.

Offline

#4 2019-02-28 14:59:22

m_vk
Member
Registered: 2019-02-20
Posts: 5

Re: XSITools error "sh: Bad block detected: 52428800"

Ok, we have confirmed that --certify-backup=yes is ok ...

...
Please wait while the (c)XSITools repository hash tables are SHA1 checked...
---------------------------------------------------------------------------------------------------------------------------
[VM0] info: SHA1 chunk hashes have been certified, 758 second taken
---------------------------------------------------------------------------------------------------------------------------
The ESXi configuration was saved to "/vmfs/volumes/backup/xsitools-repos/20190200000000/20190228064820"
---------------------------------------------------------------------------------------------------------------------------
No errors detected in backup
...

... but as soon as we use "fast", we got errors ...

...
Please wait while the (c)XSITools repository hash tables are fast checked...
---------------------------------------------------------------------------------------------------------------------------
sh: Bad block detected: 52428800
1: bad number
[VM0] error: chunk hashes fast check failed, detail:
---------------------------------------------------------------------------------------------------------------------------
The ESXi configuration was saved to "/vmfs/volumes/backup/xsitools-repos/20190200000000/20190228052910"
---------------------------------------------------------------------------------------------------------------------------
Errors detected in backup, check logs
---------------------------------------------------------------------------------------------------------------------------
...

Next backup using SHA1 check is once again without problems.

How exactly do we contact support? Maybe form at 33HOPS Contact Form ? We'd like to have fast check working as SHA1 verification takes more than 4 hours in our environment...

Last edited by m_vk (2019-02-28 15:00:28)

Offline

#5 2019-03-02 16:47:13

admin
Administrator
Registered: 2017-04-21
Posts: 2,055

Re: XSITools error "sh: Bad block detected: 52428800"

Never delete bad blocks without having thoroughly inspected the ultimate cause of the problem. Detecting a bad block is an extremely unfrequent issue. In your case the error seems to be a second derivate one, it could be a bug on either side (XSIBackup or some ESXi component).

We'll revise the --certify=fast feature on ESXi 6.7 this weekend. Nevertheless, it's a size comparison only. What is the size of the set of VMs being backed up?. We don't have that fundamental piece of information.

Offline

#6 2019-03-03 20:29:25

admin
Administrator
Registered: 2017-04-21
Posts: 2,055

Re: XSITools error "sh: Bad block detected: 52428800"

We have checked the fast XSITools block check function and it is working O.K. for us. Please, contact support so that we can offer alternatives to narrow your problem down. In any case it looks like it's something affecting just you by now.

Offline

Board footer