kafka-users mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Claudia Wegmann <c.wegm...@kasasi.de>
Subject AW: Problems with broker upgrade from 1.1.0 to 2.0.0
Date Thu, 11 Oct 2018 06:43:09 GMT
Thanks for the quick answer, Bratt.

You are right, my brokers did not shut down cleanly. The issue was not related to the update.

Thanks for the pointers!

Best,
Claudia

-----Urspr├╝ngliche Nachricht-----
Von: Brett Rann <brann@zendesk.com.INVALID> 
Gesendet: Dienstag, 9. Oktober 2018 03:26
An: Users <users@kafka.apache.org>
Betreff: Re: Problems with broker upgrade from 1.1.0 to 2.0.0

I'd

* verify you shutdown cleanly when restarting the broker
* test restarting the broker first before upgrading it (it may have been sitting on a corrupt
file for a long time and the issue is unrelated to the
upgrade)

re your questions:

1) probably not, unless it was caused by an unclean shutdown, in which case shutdown cleanly.
:)
2) probably not, since the indexes are rebuilt.

FWIW i've done a bunch of 1.1.0 -> 2.0 upgrades and haven't had this issue.
For sure have seen it in the past though.


On Tue, Oct 9, 2018 at 2:09 AM Claudia Wegmann <c.wegmann@kasasi.de> wrote:

> Dear community,
>
> I updated my kafka cluster from version 1.1.0 to 2.0.0 according to 
> the upgrade guide for rolling upgrades today. I encountered a problem 
> after starting the new broker version. The log is full of "Found a 
> corrupted index file corresponding to log file 
> /data/kafka-logs/resultGatewayDataDispatcher-22/00000000000002858226.l
> og due to Corrupt time index found, time index file
> (/data/kafka-logs/resultGatewayDataDispatcher-22/00000000000002858226.
> timeindex) has non-zero size but the last timestamp is 0 which is less 
> than the first timestamp 1538583935827}, recovering segment and 
> rebuilding index files...
> (kafka.log.Log)"
>
> Said recovering takes quite a bit of time.
>
> Now I'm not too keen to do the broker update in the production 
> environment because of the relative long processing pause.
> The questions are:
> 1.) Is there a way to avoid this error and therefore make starting the 
> new broker faster?
> 2.) Does the error imply any data loss?
>
> Any help is apricated ­čśŐ
> Best,
> Claudia
>
Mime
View raw message