the supplied seed must support extract_array()

Start with the reported neu.sce[, cell_s$cell_label] failure and the linked DelayedArray issue #112, using the supplied sessionInfo as the environment reference. Compare the working and failing RDS/container cases and inspect the HDF5Array seed involved in the extract_array() validation error. Done means a reproducible diagnosis and a confirmed fix or documented serialization/environment constraint.

I am trying to have some reproducibility using docker containers for scRNAseq analysis.
Unfortunately the data I am trying to analyze are quite large thus I need to use some aws services.
For this reason I used rds objects and a locally built container. Iuploaded everything on aws and run the large data analysis.
Every time I try to subset the singlecellexperiment object I get this issue:

> neu.sub <- neu.sce[, cell_s$cell_label]
Loading required namespace: HDF5Array
Error in validObject(.Object) : invalid class "DelayedSubset" object:
    the supplied seed must support extract_array()

The only time I didn't get the error was when I made the object inside the container, saved as rds (in the container volume mounted on my local machine), loaded it again on the container and I could subset without error.
Instead when I used a previous (almost the same object saved locally as .rds) made on my machine (not in the container) and loaded in the container, on the subsetting I got the same error.

At first I thought that it was some incompatibility of R version / package version. Now I have used the object made inside the container from my machine and load it inside the same container on aws instance and getting that error.
The data are public so I can share the code if needed for reproducibility.
Regarding the image of the container I can send (?upload?) somewhere the tar image that can be loaded (amd64) with docker load -i image.tar

Where should I start searching to debug this issue?

> sessionInfo()
R version 4.5.1 (2025-06-13)
Platform: x86_64-pc-linux-gnu
Running under: Debian GNU/Linux 13 (trixie)

Matrix products: default
BLAS:   /usr/lib/x86_64-linux-gnu/blas/libblas.so.3.12.1
LAPACK: /usr/lib/x86_64-linux-gnu/lapack/liblapack.so.3.12.1;  LAPACK version 3.12.0

locale:
[1] C

time zone: Etc/UTC
tzcode source: system (glibc)

attached base packages:
[1] stats4    stats     graphics  grDevices utils     datasets  methods
[8] base

other attached packages:
 [1] bluster_1.18.0              zellkonverter_1.18.0
 [3] scran_1.36.0                scater_1.36.0
 [5] ggplot2_4.0.0               scuttle_1.18.0
 [7] SingleCellExperiment_1.30.1 SingleR_2.10.0
 [9] SummarizedExperiment_1.38.1 Biobase_2.68.0
[11] GenomicRanges_1.60.0        GenomeInfoDb_1.44.3
[13] IRanges_2.42.0              S4Vectors_0.46.0
[15] BiocGenerics_0.54.0         generics_0.1.4
[17] MatrixGenerics_1.20.0       matrixStats_1.5.0
[19] openxlsx_4.2.8              vroom_1.6.6
[21] rlang_1.1.6                 stringr_1.5.2
[23] tibble_3.3.0                dplyr_1.1.4
[25] tidyr_1.3.1                 readr_2.1.5
[27] here_1.0.2

loaded via a namespace (and not attached):
 [1] gridExtra_2.3             magrittr_2.0.4
 [3] compiler_4.5.1            dir.expiry_1.16.0
 [5] DelayedMatrixStats_1.30.0 png_0.1-8
 [7] vctrs_0.6.5               pkgconfig_2.0.3
 [9] crayon_1.5.3              XVector_0.48.0
[11] tzdb_0.5.0                UCSC.utils_1.4.0
[13] ggbeeswarm_0.7.2          purrr_1.1.0
[15] bit_4.6.0                 beachmat_2.24.0
[17] jsonlite_2.0.0            rhdf5filters_1.20.0
[19] DelayedArray_0.34.1       Rhdf5lib_1.30.0
[21] BiocParallel_1.42.2       irlba_2.3.5.1
[23] parallel_4.5.1            cluster_2.1.8.1
[25] R6_2.6.1                  stringi_1.8.7
[27] RColorBrewer_1.1-3        limma_3.64.3
[29] reticulate_1.43.0         Rcpp_1.1.0
[31] Matrix_1.7-3              igraph_2.1.4
[33] tidyselect_1.2.1          abind_1.4-8
[35] viridis_0.6.5             codetools_0.2-20
[37] lattice_0.22-7            basilisk.utils_1.20.0
[39] withr_3.0.2               S7_0.2.0
[41] zip_2.3.3                 pillar_1.11.1
[43] filelock_1.0.3            rprojroot_2.1.1
[45] hms_1.1.3                 sparseMatrixStats_1.20.0
[47] scales_1.4.0              glue_1.8.0
[49] metapod_1.16.0            tools_4.5.1
[51] BiocNeighbors_2.2.0       ScaledMatrix_1.16.0
[53] locfit_1.5-9.12           rhdf5_2.52.1
[55] grid_4.5.1                edgeR_4.6.3
[57] GenomeInfoDbData_1.2.14   basilisk_1.20.0
[59] beeswarm_0.4.0            BiocSingular_1.24.0
[61] HDF5Array_1.36.0          vipor_0.4.7
[63] cli_3.6.5                 rsvd_1.0.5
[65] S4Arrays_1.8.1            viridisLite_0.4.2
[67] gtable_0.3.6              SparseArray_1.8.1
[69] ggrepel_0.9.6             dqrng_0.4.1
[71] farver_2.1.2              lifecycle_1.0.4
[73] h5mread_1.0.1             httr_1.4.7
[75] statmod_1.5.1             bit64_4.6.0-1

Maybe related? https://github.com/Bioconductor/DelayedArray/issues/112