Jump to content
Joe L.

cannot do non-correcting parity check after array stop/start.

9 posts in this topic Last Reply

Recommended Posts

reported by "nars":

I did found a problem, not sure if related to new gui or just a bug on latest rc's and didn't noticed it before? apparently right after starting array I'm unable to do a NOCORRECT parity check.

 

Please anyone else can try these steps and confirm this:

- stop array

- start array

- uncheck "Write corrections to parity disk"

- start parity check

isn't "Write corrections to parity disk" checkbox checked again? (and looking at syslog it is doing a 'check CORRECT')

 

Then if I stop the parity check, uncheck Write corrections to parity disk" checkbox and try again I can't reproduce it again, until stop/start array, that's why I think it only happens on first time we attempt it right after starting array.

 

Edit: I did now tested and can reproduce it on stock rc16c, on both baremetal and vmware workstation, then not related at all with the new gui... you may want to move it to issues... Also note that it doesn't happen after boot (with array auto-start), just after a stop/start as I did mention above.

Share this post


Link to post

Fixed in 'final'.  Good catch, this bug has been there a while.  Bug is first parity check following array start is forced to be the "CORRECT" variety, ignoring the checkbox.  Subsequent parity checks would honor the checkbox.

Share this post


Link to post

Can confirm this is solved on 5.0 final. Thanks again.

Share this post


Link to post

nice it is fixed, but (excuse my ignorance) is it really a prob? if i run a parity check and go for the endless time it takes - i want it to fix probs while doing it anyway, or not?

i mean, if the system is 'mumbling and bumbling' around to check parity - wouldn't it be senseless not to solve issues at the same time if possible?

 

greetz, L

Share this post


Link to post

Fixing errors as you say means supposing the data on data disks is correct and updating parity to match it, in some specific cases we may not want to do that (imagine that something is wrong with some data disk, or even some other system problem leading to data corruption, etc)... we may just want to check if all ok, i.e. if data matches parity... and it should always match unless there is some problem... that's why the option is available to do it - if you want to use it - it's an option, there was just a bug that didn't allowed to do what the option is intended to do after a stop/start array.

Share this post


Link to post

Fixing errors as you say means supposing the data on data disks is correct and updating parity to match it, in some specific cases we may not want to do that (imagine that something is wrong with some data disk, or even some other system problem leading to data corruption, etc)... we may just want to check if all ok, i.e. if data matches parity... and it should always match unless there is some problem... that's why the option is available to do it - if you want to use it - it's an option, there was just a bug that didn't allowed to do what the option is intended to do after a stop/start array.

point taken! makes sense. i consider myself also very lucky, having/ using computers since the late 80s - i never had a hdd really going down/ crashing on any of my maschines. guess i am very lucky (hope it holds up the next 30+ years). but yes i agree from this point of view.

Share this post


Link to post

Fixing errors as you say means supposing the data on data disks is correct and updating parity to match it, in some specific cases we may not want to do that (imagine that something is wrong with some data disk, or even some other system problem leading to data corruption, etc)... we may just want to check if all ok, i.e. if data matches parity... and it should always match unless there is some problem... that's why the option is available to do it - if you want to use it - it's an option, there was just a bug that didn't allowed to do what the option is intended to do after a stop/start array.

point taken! makes sense. i consider myself also very lucky, having/ using computers since the late 80s - i never had a hdd really going down/ crashing on any of my maschines. guess i am very lucky (hope it holds up the next 30+ years). but yes i agree from this point of view.

You are not lucky, you are strange then ;D unless you always sold then or something after a few years it's really strange you never got one going bad, even by age... I can say I use computers for 20 exact years and got a few already, unfortunately :( Never got one DOA though, mines just eventually died after some years... it's surely a random thing we can't predict, but they all eventually die...

Share this post


Link to post