Control: tags -1 + patch Étienne Mollier, on 2025-11-06: > If so, I consider implementing a check based on the definition > of the AUTOPKGTEST_TMP environment variable. That sounds very > hackish, but would avoid hammering on ci.d.n, while having low > risk of introducing possible performance regressions for end > users.
Said patch would look like: -------8<--------------8<--------------8<--------------8<------- Description: cap memory usage in autopkgtest context. The default calculation method to reserve memory for the java runtime environment causes excessive memory usage on ci.debian.net machines. Author: Étienne Mollier <[email protected]> Bug-Debian: https://bugs.debian.org/1120170 Forwarded: not-needed Last-Update: 2025-11-06 --- This patch header follows DEP-3: http://dep.debian.net/deps/dep3/ --- bbmap.orig/calcmem.sh +++ bbmap/calcmem.sh @@ -214,6 +214,10 @@ && [ "$RAM" -gt "2000" ] then RAM="2000" fi + # cap to 4G on shared autopkgtest resources (ci.debian.net). + if [ -n "$AUTOPKGTEST_TMP" ] && [ "$RAM" -gt "4000" ] + then RAM="4000" + fi #echo $RAM fi #local z="-Xmx${RAM}m" -------8<--------------8<--------------8<--------------8<------- I'll upload tomorrow, unless I'm told it's not a good idea to implement such check this way, while a more appropriate approach would be in order. I had hoped to pull the latest upstream version of the package while at it, but it seems something in my autopkgtest invocation lets it spin forever, and I lost a lot of time waiting for the infinite loop to finish. Anyways, let's continue tomorrow. -- .''`. Étienne Mollier <[email protected]> : :' : pgp: 8f91 b227 c7d6 f2b1 948c 8236 793c f67e 8f0d 11da `. `' sent from /dev/pts/1, please excuse my verbosity `- on air: Rausch - Drain
signature.asc
Description: PGP signature
