Network UPS Tools Memory Used ??

JonathanLee

Does anyone know the amount of system resources that NUT package uses?

netblues

Minimal for a modern system.
Judging from apcupsd15Mbytes..

51157 root 20 0 15M 3880K nanslp 0 0:14 0.00% /usr/local/sbin/apcupsd{apcupsd}

JonathanLee

Thanks for the information I had a issue with a very large blacklist that left residual code and did not perform garbage collection after removal. "Java garbage collection is the process by which Java programs perform automatic memory management." after deletion this caused memory availability size issues. I had to default the URL blacklist and it still was consuming resources until a full reimage of the firewall was performed. I had also installed nut at the same time and wanted to know if it was this package that was consuming the other 2 GBs of memory or the url blacklist. After reimage and full deletion of the large blacklist the system restored to a normal range. My url blacklist that I was testing broke Squid not Nut.

JonathanLee

@jonathanlee It is good to know what your normal system resource baseline is. We were taught that in cyber security classes to track resources issues when they are depleted unknownly to investigate issues. I found the problem.

netblues

Well, garbage collection Is something that tries to recover depleted memory resources, however, java is not used on any pfsense, as the quest for log4j told us recently.

Having said that , it is also used in python, c, c-sharp and many other modern languages.

BUT do keep in mind that a reboot of the system clears and initializes ALL ram and is effectively the ultimate garbage collector of all times.

Clearly squid was still loading the blacklist in memory. Consuming 2 gigs for a decent blacklist is more or less expected

For example, unbound consumes about 55Mbyte without any filtering lists and jumps to 715Mb with carefully selected feeds.
And when assembling final list memory usage grows even higher.

Squid, when filtering uses the same approach and is normal to consume large amounts of ram too.

JonathanLee

@netblues Yes spot on. We learned this also in Python programming class. I agree the some URLBlacklists are massive and just eat memory, makes me wonder if they use some hybrid form of Stacks, Queues, or ArrayLists. Shalla list on my Netgate 2100 Max did not have the same issues that the other blacklists are creating, Shallalist was rather large also just not as big as the others. Thanks for the reply happy Easter.