--[ Redis 2.8.14 ] Release date: 1 Sep 2014 # UPGRADE URGENCY: HIGH for Lua scripting users, the server could crash because of a bug introduced in Redis 2.8.10, otherwise LOW. LOW for Redis Sentinel. * [FIX] Don't prevent use of shared integers if maxmemory policy is non-LRU. (Salvatore Sanfilippo) * [FIX] Fail SYNC if background save child aborted due to a signal. (Yossi Gottlieb) * [FIX] Different small redis-cli fixes. (Dov Murik, Charsyam, cubicdaiya, Kashif Rasul, Jan-Erik Rediger, Matt Stancliff) * [FIX] AIX compilation fixes. (Siah Lyimo) * [FIX] A number of other smaller issues. * [FIX] Improved SIGINT handling (Matt Stancliff, Salvatore Sanfilippo) * [FIX] Use unsigned types in SDS header to raise limit to 4GB. (Matt Stancliff, Salvatore Sanfilippo) * [FIX] Handle signed/unsigned comparisons with more care around the code. (Salvatore Sanfilippo) * [FIX] Colorized test output fixed to don't change the background color. (Mariano Pérez Rodríguez) * [FIX] More Sentinel IPv6 fixes. (Eiichi Sato) * [FIX] Deny CLIENT command in scripts. (Matt Stancliff) * [FIX] Allow datasets with more than 2 billion of keys, initial work. * [FIX] Fix a Lua scripting crash by storing the length of the static argv when first allocated. (Paddy Byers) * [NEW] Pub/Sub PING. (Salvatore Sanfilippo) * [NEW] Much faster ZUNIONSTORE. (Kyle Hubert, Salvatore Sanfilippo) * [NEW] Faster ll2string() implementation. (Salvatore Sanfilippo) * [NEW] **WARNING, minor API change**: PUBSUB NUMSUB: return type modified to integer. (Matt Stancliff) * [NEW] redis-benchmark support for AUTH. (CharSyam) Tested on amd64 with the ebuild of 2.8.13.
Created attachment 385090 [details] redis-2.8.16.ebuild - bump! - fix lua path - new initd (expose restart timeout)
Created attachment 385092 [details] redis.initd-4
Created attachment 385094 [details] redis.initd-4
Thanks Johan for new init script. I just tried it and I suspect that "--retry" works in a little different way than we expect: # time /etc/init.d/redis restart * Stopping redis ... * start-stop-daemon: 1 process refused to stop [ ok ] * Starting redis ... * start-stop-daemon: /usr/sbin/redis-server is already running * Failed to start redis [ !! ] * ERROR: redis failed to start real 0m5.312s user 0m0.050s sys 0m0.180s After a few seconds: # time /etc/init.d/redis restart * Starting redis ... [ ok ] real 0m0.172s user 0m0.000s sys 0m0.030s
man page says: These options are only used for stopping daemons: -R, --retry timeout | signal/timeout The retry specification can be either a timeout in seconds or multiple sig‐nal/timeout pairs (like SIGTERM/5). Should be good, no? The mysql ebuild does something similar.
:) I've made little change: - --exec ${retries} \ + --exec ${command} \ and voila: # time /etc/init.d/redis stop * Stopping redis ... * Will stop /usr/sbin/redis-server * Will stop PID in pidfile `/var/run/redis/redis.pid' * Will stop processes of `/usr/sbin/redis-server' * Sending signal 15 to PID 30947 ... [ ok ] real 0m21.027s user 0m0.150s sys 0m0.310s "It works!"
Created attachment 385320 [details] redis.initd-4 Jesus christ, how did that end up in there? Thanks for spotting it. Updated the init attachment.
Created attachment 385360 [details] redis-2.8.17.ebuild Bump again. This is a rename of 2.8.16. # UPGRADE URGENCY: HIGH for Redis Sentinel. LOW for Redis Server (unmodified compared to 2.8.16). * [FIX] Resolved a memory leak in the hiredis library causing a memory leak in Redis Sentinel when a monitored instance or another Sentinel is unavailable. Every reconnection attempt will leak a small amount of memory, but in the long run the process can reach a considerable size.
Thanks for your patience and great work guys, it's in tree now. +*redis-2.8.17 (08 Oct 2014) + + 08 Oct 2014; Ultrabug <ultrabug@gentoo.org> +redis-2.8.17.ebuild, + +files/redis.initd-4: + version bump fix #522118 thx to Tomas Mozes, Marcin Miroslaw and Johan + Bergstrom +