Content
View differences
Updated by Varac Anero over 7 years ago
I'm using the [docker installation](https://www.openproject.org/docker/) and am upgrading from 7.2.3 to 8.1, and everything looks nice so far, with one exception.
During the first start, the textile→markdown migration is running, and there's one job which is failing:
```
Conversion: |==================================================================|
I, [2018-11-30T19:57:50.501146 #56] INFO -- : Converting WikiContents
Conversion: |
Converting items individually due to pandoc mismatch: |========================|
========E, [2018-11-30T20:01:56.704512 #56] ERROR -- : Execution of pandoc failed: Pandoc failed with code [unknown status (kill
ed?)] [Stopped=false]
Process received signal (term 6, stop none)
pandoc: internal error: Unable to commit 1048576 bytes of memory
(GHC version 8.4.3 for x86_64_alpine_linux)
Please report this as a GHC bug: http://www.haskell.org/ghc/reportabug
```
The host machine has 4G so this is not about a physical memory limit limit, but rather a haskell/pandoc bug (see https://github.com/jgm/pandoc/issues/4504).
Please make sure to ship a pandoc binary compiled with the right flags to prevent this.
Will the migration recover fromt this error - I mean, will those docs which doesn't got migrated in this failed job get processed during the next start of the container ?
If not, please implement a smart recovery mechanism from this bug.
Thanks!
During the first start, the textile→markdown migration is running, and there's one job which is failing:
```
Conversion: |==================================================================|
I, [2018-11-30T19:57:50.501146 #56] INFO -- : Converting WikiContents
Conversion: |
Converting items individually due to pandoc mismatch: |========================|
========E, [2018-11-30T20:01:56.704512 #56] ERROR -- : Execution of pandoc failed: Pandoc failed with code [unknown status (kill
ed?)] [Stopped=false]
Process received signal (term 6, stop none)
pandoc: internal error: Unable to commit 1048576 bytes of memory
(GHC version 8.4.3 for x86_64_alpine_linux)
Please report this as a GHC bug: http://www.haskell.org/ghc/reportabug
```
The host machine has 4G so this is not about a physical memory limit limit, but rather a haskell/pandoc bug (see https://github.com/jgm/pandoc/issues/4504).
Please make sure to ship a pandoc binary compiled with the right flags to prevent this.
Will the migration recover fromt this error - I mean, will those docs which doesn't got migrated in this failed job get processed during the next start of the container ?
If not, please implement a smart recovery mechanism from this bug.
Thanks!