<html><head><meta http-equiv="content-type" content="text/html; charset=utf-8"></head><body dir="auto"><div><div style="direction: inherit;">Ja jsem na tom stejne - jsem pro posileni site. Na nas mam treti repliku zaloh, plus nejaky tempy, takze to me fakt netrapi a rychlejsi sit bude urcite lepsi reseni :)</div><div style="direction: inherit;"><br></div><div style="direction: inherit;">h00ked</div><br>Odesláno z iPhonu</div><div><br>28. 7. 2016 v 14:48, Szoke Igor <<a href="mailto:igor@szoke.cz">igor@szoke.cz</a>>:<br><br></div><blockquote type="cite"><div><span>Ahoj,</span><br><span></span><br><span>za me idealne oboje :).. rychly + zalohovany. A pokud si musim vybrat, tak momentalne radeji rychlejsi s</span><br><span>prustrelnejsi zalohou.</span><br><span></span><br><span>NAS od zacatku pouzivame s vedomim nezalohovani. Mame tam >500G read-only dat (video sharing) + nejaky tempy.</span><br><span>Takze, kdyz to zdechne, tak pocitam s tim, ze to tam uploadnu znova.</span><br><span></span><br><span>Obcas muzeme delat nejaky spicky na trafic ... kdyz se sejdou lidi. Proto ta propustnost.</span><br><span></span><br><span>Toz tak,</span><br><span>-Igor</span><br><span></span><br><span></span><br><span></span><br><span></span><br><span>On 07/28/2016 02:24 PM, Pavel Snajdr wrote:</span><br><blockquote type="cite"><span>Prave bezi posledni dosyncnuti dat, odhadem za 2-3 hodiny prepneme</span><br></blockquote><blockquote type="cite"><span>mounty read-only na backuper, znovu vytvorime pool na nasboxu s</span><br></blockquote><blockquote type="cite"><span>bezpecnejsi konfiguraci.</span><br></blockquote><blockquote type="cite"><span>Pouijeme vetsi raidz2, pripadne raidz3 VDEVy, sebehnu si nejake</span><br></blockquote><blockquote type="cite"><span>benchmarky tech konfiguraci, co mne napadaji jako prijatelne a uvidime,</span><br></blockquote><blockquote type="cite"><span>co z toho vyleze, stejna konfigurace se potom casem pouzije i pro</span><br></blockquote><blockquote type="cite"><span>backuper, abychom se vyhnuli stejne situaci v budoucnu na backuperu.</span><br></blockquote><blockquote type="cite"><span></span><br></blockquote><blockquote type="cite"><span>K tomu, proc jsou v tom poolu disky se slabou redundanci, ie. 3-diskove</span><br></blockquote><blockquote type="cite"><span>RAID-Z: je to historicky dane tim, ze jak backuper, tak nasbox vznikly s</span><br></blockquote><blockquote type="cite"><span>malo disky a ZFS neumi reshape poli, pridavanim bezpecnejsich VDEVu</span><br></blockquote><blockquote type="cite"><span>bychom nic neziskali, protoze ten VDEV, ktery se rozbil, byl hned ten</span><br></blockquote><blockquote type="cite"><span>druhy v poradi, ktery tam byl uz pri prvnim vyrobeni pole.</span><br></blockquote><blockquote type="cite"><span></span><br></blockquote><blockquote type="cite"><span>Kdybychom pouzili z fleku bezpecnejsi konfiguraci, to pole by tragicky</span><br></blockquote><blockquote type="cite"><span>nestihalo na IOPS.</span><br></blockquote><blockquote type="cite"><span></span><br></blockquote><blockquote type="cite"><span>A jeste k tomu, jak NAS vubec vzniknul - to bylo tak, ze nam prebylo</span><br></blockquote><blockquote type="cite"><span>zalohovaci pole, ktere ale uz bylo male na to, aby delalo mirror zaloham</span><br></blockquote><blockquote type="cite"><span>a nechteli jsme ho nechat valet jen tak, proto jsme ho zpristupnili vsem</span><br></blockquote><blockquote type="cite"><span>a od zacatku rikali, ze neni zalohovane - mysleli jsme si, ze to pole</span><br></blockquote><blockquote type="cite"><span>vyuzijete na zalohy domacich dat a podobne, coz tedy vetsina udelala, ale...</span><br></blockquote><blockquote type="cite"><span></span><br></blockquote><blockquote type="cite"><span>Nasli se i taci, kteri pres to vsechno dali na NAS produkcni data a ted</span><br></blockquote><blockquote type="cite"><span>je cekalo velmi neprijemne prekvapeni.</span><br></blockquote><blockquote type="cite"><span></span><br></blockquote><blockquote type="cite"><span>Cili ted stojime pred rozhodnutim, jestli investovat do redundance NASu</span><br></blockquote><blockquote type="cite"><span>(a backuperu s tim), nebo jit podle puvodniho planu a upgradovat sit na</span><br></blockquote><blockquote type="cite"><span>10Gbit (coz je potreba pro lepsi debugovatelnost clusteru, kvuli kdumpu;</span><br></blockquote><blockquote type="cite"><span>a taky jsem se chystal nejak vyresit replikaci dat mezi nody).</span><br></blockquote><blockquote type="cite"><span></span><br></blockquote><blockquote type="cite"><span>Co si o tom myslite? Investovat do storage a nechat to zatim na 2Gbit</span><br></blockquote><blockquote type="cite"><span>siti (ktera je, nutno rict, sem tam, uz pekne na hrane s propustnosti)?</span><br></blockquote><blockquote type="cite"><span></span><br></blockquote><blockquote type="cite"><span>Poznamecka: prosim ujistete se, ze v odpovedi je To:</span><br></blockquote><blockquote type="cite"><span><a href="mailto:community-list@lists.vpsfree.cz">community-list@lists.vpsfree.cz</a>, na outage-list se musi prispevky</span><br></blockquote><blockquote type="cite"><span>schvalovat a mely by tam jit jenom relevantni informace o vypadcich, ne</span><br></blockquote><blockquote type="cite"><span>diskuze.</span><br></blockquote><blockquote type="cite"><span></span><br></blockquote><blockquote type="cite"><span>/snajpa</span><br></blockquote><span></span><br><span>_______________________________________________</span><br><span>Community-list mailing list</span><br><span><a href="mailto:Community-list@lists.vpsfree.cz">Community-list@lists.vpsfree.cz</a></span><br><span><a href="http://lists.vpsfree.cz/listinfo/community-list">http://lists.vpsfree.cz/listinfo/community-list</a></span><br></div></blockquote></body></html>